3D Heritage Artefacts in Education—Enhancing Attractiveness of Computer Graphics Curriculum

Abstract

Lublin University of Technology has been offering computer science studies for over 25 years. From the beginning, computer graphics have played a crucial role in the studies program. The Lab3D international efforts aimed at cultural heritage 3D digitization allowed for enriching the regular introductory course of computer graphics with elements of digitized real-life heritage, without disturbing the regular didactic process, nor the scope of the course learning outcomes. Such an activity was aimed at increasing the students’ engagement in learning and fits into the contemporary trend of including real-life cases in the course of teaching. The article presents the curriculum (including the changes introduced) together with an extensive analysis of the effect of such changes on students’ achievements. In total, 3042 grades of 1522 students from the years 2018–2024 underwent statistical analysis in order to support answering placed hypotheses related to keeping the effects positive or neutral. Luckily, students’ achievements did not worsen, and in fact significantly improved. To the authors’ knowledge, such an analysis in relation to computer graphics courses has never been published before.

1. Introduction

Lublin University of Technology (Department of Computer Science, Faculty of Electrical Engineering and Computer Science, to be precise) has been offering computer science studies for over 25 years. Many challenges have been faced over time, such as organizational aspects of studies, the COVID-19 pandemic, distanced learning and cooperation [1,2,3]. From the beginning, computer graphics have played a crucial role in the study program. The rapid development of 3D scanning techniques and international efforts aimed at popularizing the preservation of cultural heritage have opened up new opportunities for making classes more attractive for the students [4].

Lab3D is part of the Department of Computer Science at the Faculty of Electrical Engineering and Computer Science, Lublin University of Technology. The individuals who created Lab 3D were already involved in heritage digitization activities back in 2014. First, local heritage sites in eastern Poland were focused upon [5,6]. In the following years, the team’s activities internationalized to include Central Asia (e.g., Uzbekistan, tangible and intangible heritage [7,8,9]), Eastern Europe, (Romania, Cluj-Napoca region, wooden historic churches [10,11]), and even North America (United States of America, catholic churches planned for demolition [12]).

The goal of this work is to present the concepts and methods that were introduced into the computer graphics course and share the experiences gathered while conducting the classes, which were enriched with the assets of digitized cultural heritage. An attempt was also made to answer the question of whether including heritage elements into the computer graphics course negatively or positively affected students’ achievements. The analyzed period covers the years 2018–2024. The COVID-19 years were treated in a special way in this work, due to the pandemic stress, and were incomparable to the conduct of class from the other years.

Digitization activities undertaken by Lab3D allowed for introducing real-life cases into the education process, especially in areas of graphics computation, policy making, and universal design. The area most affected by changes was 3D graphics processing (modelling and texturing, creating a 3D scene and animation, rendering). It involved using models of scanned historical artefacts as the main objects, with an emphasis on presenting them on the 3D stage in accordance with the atmosphere of the object. Therefore, these did not introduce changes in the technologies used but only enabled the creation of scenes with a more interesting story. The goal of this concept was to increase the level of enthusiasm among the students when fulfilling the exercises and to improve awareness of cultural heritage topics. Moreover, elements of digitized heritage were used as practical examples provided during the other stages of the classes.

The computer graphics course was chosen as the main source of data and ideas for this article. The reason is the high quality and completeness of data during the long-term period. It has to be noted here that 3D digitization of heritage objects is a rather specific utilization of computer graphics. However, it involves many techniques from that domain and provides valuable objects that can undergo further processing. The approach for conducting the course may be perceived as the implementation of problem- and project-based learning [13].

The following research hypotheses were formulated:

H1. 

Introduction of digitized cultural heritage elements to the computer graphics laboratory has a positive effect, which translates into better grades obtained by students at the end of laboratory classes.

H2. 

Exam grades obtained by students at the end of computer graphics lectures are better than before introducing digitized cultural heritage elements to the computer graphics course.

H3. 

Introduction of digitized cultural heritage elements to the computer graphics laboratory has a positive effect (e.g., better memorability or encouragement to dive into theoretical aspects), which translates into better exam grades obtained by students at the end of lectures.

H4. 

Introduction of digitized cultural heritage elements to the computer graphics course has a positive effect (e.g., better memorability or encouragement), which translates into a significantly lower number of negative grades given to students during the first attempts to pass.

H5. 

Due to the extraordinary circumstances of education, results obtained by students during the COVID-19 pandemic are different than after it finished.

H6. 

Introduction of digitized cultural heritage elements to the computer graphics course makes it more interesting for the participating students.

The authors believe that the strength of the article is not only the proposal of a computer graphics curriculum that has been tested over many years of practice and takes into account the richness of the digital cultural heritage assets, but also the impact verification of introducing such content on students’ results. The rarely encountered large research sample and long duration of observation cannot be underestimated either. Therefore, this article is of particular interest and value to researchers and teachers who are searching for proven curriculum that can be adopted, or who want to enrich their syllabi with elements of digitized cultural heritage, but are afraid of the negative effects of such changes.

2. Literature Review

The literature review focused on finding works of other authors on the subject of the computer graphics curriculum proposal for students in technical fields (especially computer science), whose curriculum would be expanded with elements of digitized cultural heritage. For this purpose, the bibliographic databases WOS, Scopus and Google Scholar were searched for the occurrences of the following three keywords: computer graphics, curriculum, heritage. The word “curriculum” was finally replaced with “education”, which returned a bigger (and a lot more reasonable) set of results. Each time, the equivalent of the logical operator “and” was used between the individual keywords. No works by other authors were found that would focus on this topic as closely as this article.

One of the most notable works found was [14], which studied how computer graphics is taught and proposed a course on 2D graphics and image processing as an alternative to the traditional 3D graphics course. It was based on an analysis of more than 70 computer science curricula from five countries. It also references other important works, stating what computer graphics courses should look like in terms of content. What is more, Cunningham [15] discussed how to meet the European Bologna requirements. In general, the first 10 years of the 2000s can be seen as the time of the most lively discussions. Balreira et al. [16] asked the question, “What are we teaching in Introduction to Computer Graphics?”. They surveyed 20 courses from higher-level educational institutions around the world on the topics that are taught. Mosendz et al. [17] investigated the impact of experimental methods for teaching computer graphics and design on fostering the creative potential of higher education students. Zhang et al. [18] shared their experiences on constructing a computer graphics course that integrates technology and art.

At this point, several conclusions can be drawn. The first is that the curriculum, used at the Lublin University of Technology, complies with the frames set up by the above-mentioned works. The second is that a significant part of the works found has one or more of the following shortcomings: a detailed computer graphics curriculum not presented, an aspect of heritage enrichment missing, and long-term validation missing. Contemporary works focus more on the computer graphics-chosen areas only and sophisticated teaching methods, like teaching computer graphics in virtual reality [19,20,21] or dealing with remote teaching intricacies [22,23], which has become typical for modern education.

Adding another search phrase element, “heritage”, gave a long list of works thematically related to the teaching of computer graphics and cultural heritage. Some of them analyzed the need to set up a computer graphics course, specifically crafted for non-engineering studies purposes, namely art studies [24], heritage conservation [25], and museology [26]. Undergoing such craft usually indicates significant shortcomings in covering the entirety of mandatory topics, especially from the computer science students’ perspective, which makes these works incomparable to the article. Some works focused solely on a chosen computer graphics aspect, e.g., 3D rendering [27], 3D graphics in general [28], photogrammetry [29], augmented reality [30], computer-aided design [31], or artistic values [32,33]. Others focused on particular narrow applications of materials obtained via computer graphics tools, e.g., making and printing a 3D model or making a game that will achieve a certain didactic effect [5]; using virtual reality scenes in heritage education [34]; aiding historic reconstruction [35]; and more.

What is more, in the context of using heritage in education, few works seem to be notable. In [36], dynamics between education and heritage were discussed. In [37], the saturation of high school curricula in Turkey with cultural heritage topics was analyzed. In [38], teachers’ points of view on heritage presence in the teaching–learning process and the use of cultural heritage in the classroom were discussed. In [39], the authors argue that the design-thinking approach varies across disciplines, including cultural heritage, language, and technology.

In the case of mixing computer graphics teaching and cultural heritage, other authors primarily attempt to use ICT technologies in teaching some theory and practical skills related to the conservation, preservation and promotion of cultural heritage (e.g., [20,29,40]), and related topics. The article focuses on something completely different—a course intended to introduce computer scientists to computer graphics, which is enriched with elements of digitized cultural heritage. As a result, a full-scale curriculum, addressed to students of computer science and related studies, is presented. It takes into account contemporary trends, as well as the experience of IEEE/ACM [41,42]. Next, the article presents some of the heritage-related materials enriching the course (a bit more of them can be seen in [4]). Significant effort is needed to select these materials to address the variety of operations, transformations and applications typical for computer graphics, satisfying the necessary level of diversity of content presented in such courses. The enrichment was performed while maintaining the original identity of the computer graphics course to provide additional motivation and interest to students, as well as to promote the cultural heritage and digitization activities of the Department of Computer Science at the Lublin University of Technology.

The initial effort to describe the above-mentioned initiative was undertaken in 2021 in [4]. Unfortunately, it is incomplete in many aspects, which motivated us to write the current article. The main problem of the previous article lies in it being shaped by the size limits imposed by the specificity of a post-conference publication. There was no space for a full presentation of the curriculum, the way of enriching it with cultural heritage content, and data analyses. Another problem was the coincidence of the COVID-19 pandemic with the curriculum enrichment. Well-established and common methods of teaching and checking learning outcomes became disrupted (remote equivalents emerged), people lived stressful lives, and the level of control over the curriculum execution was significantly decreased; thus, it was uncertain if the collected data were trustworthy. Therefore, a major change in the current article is the inclusion of data from 2022–2024 in the analysis, when conditions of the curriculum execution were fully comparable to 2018–2019. Nevertheless, COVID-time results might be perceived as interesting; thus, they were included in analyses as a separate period. Other differences are more advanced statistical methods used to analyze the data and more diverse hypotheses put forward regarding the relationships between more detailed data. In the previous article, only the average of the final grades from lectures and laboratory classes was statistically analyzed. Finally, there is enough space this time to discuss controversial aspects of the research.

3. Materials and Methods

This section introduces the computer graphics curriculum applied by the Department of Computer Science (Lublin University of Technology) for the purpose of computer science studies. The authors explain how it was enriched by elements of digitized cultural heritage. Next, the procedure, statistical methods, and threats to the study are presented.

3.1. Computer Graphics Curriculum

Teaching computer graphics involves the creation and processing of graphical content by students. In addition, it is a compromise between the thorough knowledge of the theoretical and algorithmic foundations of computer graphics and the knowledge of a wide range of its applications [16]. It is hard to imagine computer science studies without a computer graphics course. Naturally, each university has its individual approach to this matter. The Department of Computer Science at the Lublin University of Technology, when shaping the contemporary study program in the field of computer science, decided to take inspiration from the experience of IEEE/ACM in this area. It has influenced the shaping of the frames of the curriculum, presented in this section, which was used in the period covered by this research.

The computer graphics course is conducted during the fourth semester of the first degree in computer science, in both full-time and extramural studies. For full-time studies, 30 h of lectures and 30 h of laboratory classes are planned. The time of individual work of a typical student is defined as another 65 h. For extramural studies, 15 h of laboratory classes and lectures are planned, respectively, as well as another 95 h of a student’s individual work. Student’s individual work is related to the course, performed in any place, but outside the hours planned for lectures and laboratory classes. Of course, depending on the skill level, curiosity and commitment, students may exceed (or not) the time planned for individual work. Student workload is the only difference between the curriculum for full-time and extramural studies. The number of ECTS points was set at five in each case. The student’s workload summary is presented in

Table 1. Student’s workload summary.

Form of Activity	Average Number of Hours—Full-Time Studies	Average Number of Hours—Extramural Studies
Classes	60	30
- lectures	30	15
- laboratories	30	15
Individual work	65	95
- making projects	20	30
- preparing for laboratory classes	30	40
- preparing for the exam	15	25
Total	125	125

.

Teaching methods include lectures involving multimedia presentation and work in a computer laboratory. Lectures end with a written exam grade. Laboratory classes end with an assessment grade, which takes into account completing laboratory tasks and doing a project. Each time, the student is granted three attempts (chances) to receive a positive grade. After getting a positive grade, no further attempts are granted. The student cannot take the exam before obtaining a positive assessment grade.

The computer graphics course objectives were defined as (a) familiarizing students with basic topics of computer graphics and methods of graphics processing; (b) acquiring skills in processing of 2D, 3D, raster and vector graphics by students. In the preliminary requirements in terms of knowledge, skills and other competencies, the important role of knowledge of mathematics for computer scientists was indicated.

The tendency in the Polish education system is to briefly define learning outcomes—a small number of outcomes reflecting the essence of a course. The outcomes, in turn, translate into topics to be discussed during lectures and laboratory classes. The outcomes concern three areas—knowledge, skills and social competences. This approach fits into the European educational framework.

In the case of the computer graphics course, in terms of knowledge, a student:

• has structured, theoretically based general knowledge covering the topics of processing 2D and 3D computer graphics,
• knows and understands the methods, techniques and basic software for processing computer graphics.

In terms of skills, a student:

• is able to correctly interpret and use notations and methods used in computer graphics,
• is able to perform a critical comparative analysis of existing methods for obtaining the desired effect of transforming a raster image or a vector object,
• is able to select and apply in practice the appropriate method for carrying out a simple task in the field of 2D and 3D computer graphics processing.

In terms of social competences, a student is able to assess his/her competences in carrying out the task of analyzing and transforming computer graphics.

After transferring the learning outcomes to the class program, it was decided that the lectures should cover the following topics:

• Basic concepts of 2D graphics.
• A review of tools and environments facilitating the processing of raster graphics.
• Basic raster operations, context-free and context-based transformations, and interpolation.
• Raster image compression.
• A review of tools and environments facilitating the processing of 2D vector graphics, geometric transformations of 2D objects, and drawing curves.
• Basic concepts of 3D graphics, types and construction of 3D objects, construction methods, and hierarchization of 3D objects.
• Geometric transformations of 3D objects.
• A review of tools and environments facilitating the processing of 3D graphics.
• Operations on 3D meshes.
• 3D scene visualization, 3D image projection, basics of calculating the lighting of 3D scenes, lighting and shading models
• Basics of raster and vector animation, types of animation.

Topics for the laboratory classes were formulated as follows:

• Getting to know commonly available tools for editing raster graphics. Simple processing of raster images: correction of image parameters, scaling, rotation, cropping, and use of filters provided by an editing application.
• Dividing an image into layers, modifying 2D image slices, image corrections using area tools, retouching and assembly.
• Getting to know an environment supporting raster image processing, loading a raster image for processing, and building algorithms implementing typical filters (brightness, contrast correction, scaling with interpolation, etc.).
• Getting to know commonly available tools for editing vector graphics. Building simple vector images, techniques for processing vector objects. Combining vector and raster images.
• Using an environment for defining and processing vector images. Building algorithms implementing typical vector transformations.
• Getting to know tools for editing 3D graphics. Creating basic 3D objects—parametric and mesh models, using primitives and extrusion.
• Creating complex objects using CSG methods, editing and modifying a model mesh. Building a 3D object based on a pattern.
• Building 3D scenes. Basics of using scripting languages to automate the construction and transformation of 3D scenes.
• Simple application of textures to 3D objects, UV distribution.
• Getting to know tools for building raster and vector animations.
• Creating animated objects and building simple motion animations.

3.2. Enrichment with Elements of Cultural Heritage

The curriculum of the computer graphics course does not include aspects of cultural heritage. It was therefore not recommended to directly include issues of the digitization of historical objects in the curriculum content. However, cultural heritage objects and models created on their basis served as examples of discussed aspects and methods in the relevant areas of graphics presented in the lecture (Figure 1).

An exam, at the end of lectures, was conducted in the same form as a single-choice test during the period covered by the article. Changes to the test questions were limited to modifying the content of the questions and answers within the same scope of issues. They did not include any elements of cultural heritage, so any knowledge of this subject by the student had no influence on the result achieved.

Laboratory classes cover three main areas:

• Raster graphics—from basic non-contextual editing to image assembly.
• Vector graphics—the basics of building vector elements, creating simple and advanced shapes, filling with textures, etc.
• 3D graphics—building and texturing simple models, setting lighting, creating and animating scenes, while utilizing ready-made assets. Knowledge and skills from previous stages are taken into account.

Such a division assumes the expansion of the student’s knowledge and skills from the basics to the ability to build and manage a 3D scene containing self-created elements and ready-made models, along with the ability to create basic animations of scene elements. As the students progress in education, they use increasingly complex constructs, built from elements whose concepts they learned at earlier stages. This facilitates the understanding of the relationships between the components of 3D graphic models.

After the stage of performing tasks according to the instructions, a design task begins, requiring a student to come up with an idea of a virtual world concept, which will be represented as a 3D scene. Depending on the students’ predispositions, this may make it easier or more difficult to demonstrate the acquired knowledge and skills (having or not having an idea for an interesting scene).

The possibility of using models of historical objects in this task was a natural enhancement of this stage of learning. Imposing a 3D model of a specific historical object as the central element of the scene made it easier to start designing its layout (Figure 2). The models made available to students come from the Afrasiab Museum of Samarkand, Uzbekistan [43], Archeological Museum of Khoja Akhmet Yassawi International Kazakh-Turkish University, Turkestan, Kazakhstan, and the Scientific-Experimental Museum—Laboratory of the Samarkand State University in Uzbekistan [44] and present utility objects from the 13–16th century AC.

The number of objects was enough to keep objects unique for each student in one year. The original artifacts were scanned using the structural light technique with the Artec EVA scanner (https://www.artec3d.com/). A typical outcome was a 3D mesh model between 100,000 and 500,000 triangles. To avoid a situation where the complexity of a model would influence the processing time and thus introduce inequality between the computation time required, all the models were simplified to fewer than 50,000 triangles (Figure 3). The simplified models were randomly distributed among students.

The 3D arrangements produced by students were assessed on their merits in terms of the techniques used and the application of principles in 3D scene modeling. The adequacy of the scene produced to the historical background of the main object used was not taken into account. In this way, the influence of the student’s knowledge of cultural heritage on the assessment of their skills in computer graphics was avoided.

The 3D modeling section of the course included the task of building and animating a scene centered around a base model that was to play the leading role. Students were given individual models drawn from an available pool. Until 2020, this pool included typical everyday objects such as office elements, smaller furniture, dishes and other kitchen equipment, etc. The emphasis was on these elements presenting similar graphic complexity (similar size and density of the triangle mesh, similar texture quality). The utilitarian features of the object were deliberately varied; students were required to build a scene that was graphically correct and close to the actual image (correct lighting, shadows, backgrounds, etc.). No specific subject matter was imposed on the scene so that the student could present the arrangement that they think they can build best. The change introduced in 2020 was not intended to disrupt the way this final work was created or its assessment. It was decided upon to make the process of building the scene more attractive by exchanging models of typical utility objects for more “interesting” historical objects, characterized by a certain mystery. The idea, “What could this object be used for and how could I present it in 3D?” was intended by the instructors to strengthen the students’ desire to create more ambitious works.

As a result, changes in the curriculum and content of the classes were avoided while maintaining the same scope of teaching. Only the “teaching aids” used during the implemented projects were subject to exchange.

3.3. Procedure

The goal is to answer the question of whether including heritage elements into the computer graphics course negatively or positively affected students’ achievements. The analyzed period covered the previous 8 years. The COVID-19 years were treated in a special way in this work, due to the pandemic stress, and were incomparable to other years’ way of conducting the classes.

3.3.1. Analyzed Periods

The achievements of students from the years 2018–2024, in which the computer graphics course was conducted according to the curriculum described in Section 3.1, were analyzed. These years were divided into three periods:

• 2018–2019—called “Before CH”—these years concern the computer graphics course execution without elements of digitized cultural heritage. These are also the years when lectures and laboratory classes, as well as exams and assessments, were conducted on-site at the university in direct contact with the teacher in the full scope of time planned for laboratory classes and lectures.
• 2020–2021—called “COVID-19”—these years concern the computer graphics course execution in a new way, i.e., with the inclusion of digitized cultural heritage elements in the course. Unfortunately, they overlap with the COVID-19 pandemic, when lectures and laboratories were held remotely, as were exams and assessments. There was no direct contact with the teacher throughout the time planned for classes. This was also accompanied by general stress related to the direction of the pandemic’s development.
• 2022–2024—called “After CH”—these years concern the computer graphics course execution with the inclusion of digitized cultural heritage elements. Again, lectures and laboratory classes, as well as exams and assessments, were conducted on-site at the university in direct contact with the teacher in the full scope of time planned for laboratory classes and lectures.

“COVID-19” is an interesting period because the relationship between a student and a teacher was disrupted. The teacher had to play a supervisory and supportive role using not entirely perfect methods of remote work, such as remote exams on the Moodle platform, or discussing/grading projects and conducting regular teaching activities via videoconferencing tools, MS Teams, Moodle and e-mail. It was an attempt to remotely achieve learning outcomes that had previously been achieved in person. Often, both sides were accompanied by stress related to the restrictions of the pandemic or the uncertainty of tomorrow. In connection with the above, it is interesting how much the results and the entire field of experiment were disrupted compared to the “After CH” period, which is comparable to the “Before CH” period in terms of the teaching model and the social situation.

3.3.2. Grading System

In order to determine the impact of introducing digitized cultural heritage elements into the computer graphics course, it was decided upon to use students’ final grades from lectures and laboratory classes. In accordance with the intention of the curriculum, such a grade reflects the degree to which the learning outcomes specified for the subject have been achieved. The acronym LOR (learning outcomes rate) is introduced here, which refers to the percentage level of learning outcomes achieved by a student. The Polish academic grading system uses the following grades: 2.0 (negative grade, LOR below 51%), 3.0 (the lowest positive grade, means good enough to pass, LOR equals 51–60%), 3.5 (positive grade, LOR equals 61–70%), 4.0 (middle positive grade, means good, LOR equals 71–80%), 4.5 (positive grade, LOR equals 81–90%), 5.0 (the highest positive grade, means very good, LOR above 91%). In the general case, the percentage limits of LOR for positive grades are rather conventional, but they are often close to those presented.

It was decided to analyze the grades of full-time students only. The attendees of extramural studies were not included due to the significant differences in their background negatively influenced the comparability of data. The most important factors were classes being half shorter, and significantly more hours of non-directly supervised own work. These resulted in a significantly lower degree of the class’s environment monitoring and reliability of the gathered data.

3.3.3. Data Acquisition

Unfortunately, only a metric, in the form of exam and assessment grades, has remained available from the perspective of years—8 years ago, the need for regular surveys in this area was not foreseen. The data obtained from electronic protocols from the virtual dean’s office contained course data (including course name, student group number, academic year), student data (including first name, last name, student identifier, grades and dates of their acquisition), and teacher data (including first name, last name, academic title). These data were anonymized and did not contain demographic data such as sex, etc. Student identifiers and names of teachers (responsible for the particular groups of students) were used only to identify, code and statistically explore the relationships between the data. The results presented in the article do not include student data at all, only aggregated lists of paired grades (pairing means assigning to a student his exam and assessment grade). No teachers’ data were revealed—they were just coded from T1 to T6, where T stands for teacher.

Due to the large number of computer science students at the Lublin University of Technology, they are divided into lecture and laboratory groups. The lecture group is assigned to one teacher, and it contains all students of a given year. There are also many laboratory groups, each containing about 15 students. These groups are divided among many teachers, but with the assumption that a given group is led by one teacher only. There was a favorable coincidence whereby the lecture for all the analyzed years was led by the same teacher, the co-author of this article. He was also responsible for the shape of the curriculum and enforced the compliance of the other teachers involved in the implementation of the course in the field of laboratory classes. The same teacher also led a significant number of laboratory groups every year. The laboratory classes were led by a total of six teachers, five of whom performed it every year, and to a significant extent. Only the role of the sixth teacher was episodic and concerned with one laboratory group led in one year.

At last, each year, a different group of students took the computer graphics course, except for those who repeated the course due to unsatisfactory learning outcomes. These very rare cases were not excluded from the analyses, although their special status was not taken into account.

3.3.4. Results Comparability

The course, which is an introduction to computer graphics, due to its introductory role, has a stable scope of presented content in the long term. During the years 2018–2024, the same curriculum was used. In 2020, giving regular 3D models depicting everyday things was abandoned, and students started to receive 3D models as a result of cultural heritage digitization works. Years 2018–2019 and 2022–2024 are comparable, as the experimental setup was fully reproducible. Years 2020–2021 are COVID-19 years and they are not comparable due to the specificity of conducting classes remotely and in stress. They were included in the analysis as a curiosity, because the results do not occupy too much space, while being interesting for the reader.

Lectures and exams were conducted by the same teacher based on the curriculum presented in Section 3.1. The method of conducting them also did not change. In the case of laboratory classes, the analogous method of conducting classes by each teacher was enforced by the lecturer (the person responsible for the whole course). Throughout all analyzed years, students worked on individual assignments, with the advisory role of a teacher. The same instructions and laboratory materials were used by the teachers during a particular year. This means that when sets of 3D models changed in 2020, it happened for everybody at the same time.

During the exam and assessment, the student’s independence was monitored. No auxiliary sources of knowledge could be used during the exam. Cheating was forbidden, and when detected, students received negative grades. Projects and tasks were questioned live, asking for explanations and modifications, in order to verify the student’s assumed acquisition of skills during the implementation of a project or stage task.

Teachers were aligned in grading and delivery. There were commonalities in the years 2018–2024: the same criteria for giving the grades (checklist of what students’ outcomes should include and when), the same grading scale, the same curriculum, the same teachers (excluding the already unreliable “COVID-19” period).

The course modification, by including certain digitized cultural heritage elements, was not intended to change the way classes were conducted and evaluated, but only to make them more attractive by introducing interesting heritage-related materials that would be subject to operations typical for computer graphics. This is another factor supporting the comparability of educational results obtained over the analyzed years.

The Lublin University of Technology imposes strict and uniform standards and policies of grading students and quality of teaching in general. Nevertheless, as usual, a few threats to the study must be mentioned. They are discussed in Section 3.5.

3.3.5. Software and Hardware

For the course purposes, equipment with sufficient computing power to smoothly complete tasks was used. The requirement for the equipment was to fulfill tasks in software such as The Gimp, Inkscape, Processing and Blender during the laboratory without significant delays. Tasks and materials remained of the same complexity during the whole period. The hardware used was either university equipment or a student’s private equipment, when the student requested it and the teacher agreed. We do not provide the exact hardware parameters because they have changed over the years, adapting to technological advancements. Similarly, the software used has evolved along with the IT market, but attention was paid to the fact that version upgrades should not disturb the learning process and allow students to focus directly on implementing the course program assumptions and acquiring specific learning outcomes throughout the whole curriculum.

3.3.6. Procedure Summary

A simplified procedure used during the research can be written as follows:

• In 2018–2019, the computer graphics course was conducted according to the curriculum presented in Section 3.1, but without enriching the course with elements of digitized cultural heritage. Students first passed laboratory classes, then lectures, obtaining two kinds of grades.
• In 2020–2024, the computer graphics course was conducted according to the curriculum presented in Section 3.1, but after enriching the course with elements of digitized cultural heritage. Students first passed laboratory classes, then lectures, obtaining two kinds of grades.
• Before giving the grades, the independence of a student’s work and the actual acquisition of specific knowledge and skills were actively verified.
• The grades obtained by students (research data) in 2018–2024 were exported from the dean’s office system.
• The exported data were totally anonymized, leaving only the artificial student’s number, teacher’s number, and student’s grades. It was ensured to preserve and maintain the relationship between a student, his teachers and his grades for the purposes of statistical analyses.
• Statistical analyses of the research data were performed according to the methods described in Section 3.4.

3.4. Statistical Methods

Box plots, histograms and scatter plots were used to graphically present the data. Due to the ordinal nature of the analyzed variables, nonparametric tests were used. When comparing at least three independent groups, the Kruskal–Wallis test was used, and as a post-hoc, Dunn’s test with the Benjamini–Hochberg adjustment method was used to control the false discovery rate. The asymptotic significance test of the Spearman correlation coefficient was used for correlations between variables. The chi-square test was used to compare two nominal variables. In the case of a significant relationship, pairwise comparison of proportions with the Benjamini–Hochberg adjustment method, to control the false discovery rate, was followed. Statistical significance was set at p < 0.05. The results were analyzed using R 4.4.2 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria).

3.5. Threats to the Study

The research and analyses presented in this article were conducted over a multi-year period, with the changing state of technology and society. The participants of the described computer graphics course change from year to year, and the analyzed data are generated by different people (teachers). In addition, the analysis of the grades themselves is a somewhat simplified, black-box approach. The authors are aware of the resulting risks, but they also believe that, in addition to sharing their experiences, it is possible to detect certain general regularities (patterns) that would be useful for the academic community and enthusiasts of cultural heritage. This is facilitated by the mathematical apparatus used, a very large population size, and a long time of observation. Unfortunately, it is not possible to avoid all risks. The more interesting ones are described in the following paragraphs.

Due to long-term observation (since 2018), we are exposed to the effect of the generation shift. This can manifest itself in a positive way, e.g., a greater degree of familiarity with technology, resulting in greater ease in solving tasks set for students. It can also manifest itself in a negative way, e.g., through a change in priorities and philosophy of life, manifested by reduced motivation to actively participate in classes and gain achievements. Unfortunately, the authors are not able to counteract this threat. Fortunately, the effect of the generation shift should not be radical over an 8-year period.

Advancement in tools used for teaching (software and hardware) cannot be avoided. Sticking to the same versions of software and hardware from 2018 would ultimately discourage students, presenting the course as archaic and out of touch with the present. Hardware development can be a neutral or motivating factor for students because, with more efficient hardware, they can create more realistic and complex things. On the other hand, in terms of software, either programs were updated to newer versions or, as a last resort, a substitute with similar characteristics was sought, which would allow achieving learning outcomes in a similar way as before. Substitutes with an overwhelming wealth of functions not directly related to the needs of laboratory classes were avoided. The aforementioned relatively low potential for change, related to the specific role of the introductory course in computer graphics, is a facilitation.

In order to determine the impact (positive, neutral or negative) of introducing elements of digitized cultural heritage to the computer graphics course, it was decided to use students’ final grades from lectures and laboratory classes. In accordance with the intention of the curriculum, such a grade reflects the degree to which the learning outcomes specified for the course have been achieved. In some sense, this method is imperfect because it does not provide insight into the full process that led to the student achieving his grade. Unfortunately, there were no systematic surveys examining students’ attitudes to the introduced changes, as well as other aspects probably useful for this work, because the authors did not anticipate in 2018 that they would come up with the idea of such an article. Insight into subjective feelings was lost, but we believe that the scale of the research sample, the long-term observations, as well as the proposal of a curriculum verified by many years of practice, still make this article valuable. At the same time, we have to accept the fact that we lose sight of opinions, emotions and other conditions, and the evaluation becomes something like a simplified black box.

Fortunately, the problem with missing data does not exist. It was possible to obtain high-count data, with all data records having a complete set of associated values (with one exception described later). In each year, a complete set of grades for the entire year of full-time computer science studies was obtained. It is not problematic whether students have submitted all the stage tasks within the course, because teachers have given a complete set of grades, adequate to the student’s progress.

The large number of good grades may attract attention. The answer is that the students were simply excellent and committed. Computer science studies are currently very popular, offering the prospect of prestigious employment and a comfortable life. As a result, there are many more applicants than places offered, which results in accepting top students, which is also visible in the university’s recruitment data.

Finally, a typical problem of this type of research appears, which is the biased approach of the teacher during a student evaluation. In the case of exam grades, there is no such problem, because each element of tested knowledge is clearly bulleted, described and assessed binarily. The content of the examination questions evenly covers the topics listed in the curriculum, translating into the content presented during the lecture. In the case of laboratory classes, there is room for a certain level of teacher subjectivity, although an attempt was made to eliminate it by introducing a checklist to check whether the student works contain the appropriate elements implemented to the appropriate extent. In addition, the situation of potential bias is counteracted by the stable number of five teachers who led laboratory groups of fairly balanced numbers each year. Finally, the intent is to utilize statistical methods to discover the general trends in data related to the introduction of digitized heritage elements into the computer graphics course.

4. Results

It was possible to gather grades of 1522 computer science students at the Lublin University of Technology, from the years 2018–2024: 1520 exam grades and 1522 assessment grades. Two of the students were retaking only the laboratory classes (the dean’s office did not assign them to the lectures), which explains the difference in numbers. On average, a student’s age was between 18 and 25 years old (it results from the specificity of the Polish higher-education system), although there is a rare chance that single students may be older than that. Unfortunately, the authors were not granted the ability to export data containing the exact age of respondents.

Table 2. Descriptive statistics of grades.

	Year	n	Mean		Median (Q1; Q3)
			Lab	Lecture	Lab	Lecture
Before CH	2018	179	4.33	4.14	4.5 (4.0; 5.0)	4.5 (3.5; 5.0)
	2019	188	4.05	3.98	4.0 (3.5; 4.5)	4.5 (3.5; 4.5)
	Total	367	4.18	4.06	4.5 (4.0; 4.5)	4.5 (3.5; 4.5)
COVID-19	2020	206	4.22	4.08	4.5 (4.0; 5.0)	4.5 (3.5; 5.0)
	2021	234 *	4.35	4.17	4.5 (4.0; 5.0)	4.0 (3.5; 5.0)
	Total	440 **	4.29	4.13	4.5 (4.0; 5.0)	4.5 (3.5; 5.0)
After CH	2022	285	4.49	4.52	4.5 (4.0; 5.0)	4.5 (4.5; 5.0)
	2023	185	4.61	4.57	5.0 (4.5; 5.0)	4.5 (4.5; 5.0)
	2024	245	4.48	4.45	4.5 (4.0; 5.0)	4.5 (4.0; 5.0)
	Total	715	4.52	4.51	4.5 (4.5; 5.0)	4.5 (4.5; 5.0)

* 232 people were assessed from the lecture this year. ** 438 people were assessed from the lecture during the “COVID-19” period.

holds basic descriptive statistics (number of students—“n”, mean, median, quartiles—“Q1”, “Q3”) for individual years and is also divided into periods: “Before CH”, “COVID-19”, “After CH”. The “Lab” column holds values for the laboratory classes, and the “Lecture” column for the lectures. The average assessment grade (respectively exam grade) has gradually increased over the years, from approximately 4.18 (respectively 4.06) in the “Before CH” period, through approximately 4.29 (respectively 4.13) during the pandemic period, to 4.52 (respectively 4.51) in the “After CH” period. If a student receives a negative grade on the first attempt, the average of all attempts is used for analysis.

A strong, statistically significant correlation was found between exam (lecture) and assessment (laboratory) grades.

Table 3. Correlation between exam and assessment grades in individual years and in periods.

	Year	Rho	p-Value
Before CH	2018	0.89	<0.001
	2019	0.81	<0.001
	Total	0.84	<0.001
COVID-19	2020	0.89	<0.001
	2021	0.91	<0.001
	Total	0.9	<0.001
After CH	2022	0.98	<0.001
	2023	0.85	<0.001
	2024	0.95	<0.001
	Total	0.94	<0.001
Total	2018–2024	0.92	<0.001

shows the Spearman correlation coefficient and its p-value calculated overall, by year, and also divided into investigated periods. The higher the assessment grade a student received, the better he performed on the exam. The strength of correlation increased from 0.84 (“Before CH” period) through 0.9 (the “COVID-19” period) to 0.92 (“After CH” period). Figure 4 shows the scatter plot of Spearman correlation coefficient values in individual years along with the trend line.

In order to check whether there are differences in the distribution of exam grades depending on the period (“Before CH”, “COVID-19”, “After CH”), the Kruskal–Wallis test was performed (null hypothesis stating that there are no differences between all periods). Statistically significant differences were found (p < 0.001); thus, in the next step, a test comparing pairs of periods was performed to find periods that differed. The post-hoc Dunn’s test (null hypothesis stating that there are no differences between the two compared periods) showed significant differences between all groups. Namely, between the periods “Before CH” and “COVID-19” (p = 0.033), the periods “After CH” and “COVID-19” (p < 0.001)—shown in Figure 5, and the periods “Before CH” and “After CH” (p < 0.001). Analysis of the Dunn test results and data distribution showed that exam grades given during the “COVID-19” period were better than “Before CH”, and worse compared to “After CH”.

Assessment grades were analyzed in a similar way to exam grades. The null hypothesis stated that there are no differences in assessment grades between all periods. Figure 6 shows statistically significant differences (Kruskal–Wallis test, p < 0.001) in assessment grades depending on whether the grade was given in a particular period. The post-hoc Dunn’s test (the null hypothesis stated that there are no differences in assessment grades between the two compared periods) showed significant differences between all groups. Namely, significant differences in assessment grades were found between the “Before CH” and “COVID-19” periods (p < 0.001) and between the “COVID-19” and “After CH” periods (p < 0.001). The periods “Before CH” and “After CH” also differed significantly (p < 0.001). Analysis of the Dunn’s test results and data distribution showed that assessment grades given during the “COVID-19” period were better than “Before CH” and worse compared to “After CH”.

Before introducing cultural heritage elements into the course, 75% of students received a grade of 4.5 or less. During the pandemic, half of the students received a grade of 4.5 or more, with the top 25% receiving a grade of 5. In the last period, 75% of students received a grade of 4.5 or more. The pattern was similar for both labs and lectures (see Figure 5 and Figure 6).

Table 4. Number and percentage of failing grades compared to other grades in the periods “Before CH”, “COVID-19”, and “After CH”—laboratory.

	Failing Grades	Other Grades	Total
Before CH	13 (3.54%)	354 (96.46%)	367 (100%)
COVID-19	19 (4.32%)	421 (95.68%)	440 (100%)
After CH	17 (2.38%)	698 (97.62%)	715 (100%)

(respectively

Table 5. Number and percentage of failing grades compared to other grades in the periods “Before CH”, “COVID-19”, and “After CH”—lecture.

	Failing Grades	Other Grades	Total
Before CH	27 (7.36%)	340 (92.64%)	367 (100%)
COVID-19	15 (3.42%)	423 (96.58%)	438 (100%)
After CH	12 (1.68%)	703 (98.32%)	715 (100%)

) contains the number and percentage of students failing the first attempt to pass laboratory classes (respectively lectures). The hypothesis that there is no relationship between failing the first attempt to pass laboratory classes and the period in which the assessment was carried out was tested. After performing the chi-squared test, there was no significant relationship between the percentage of people failing the first attempt and the period in which the assessment was carried out in the case of assessment grades (p = 0.178). However, a significant relationship was found in the case of exam grades (p < 0.001). After performing the pairwise comparison of proportions (the null hypothesis stated that there is no significant difference in the percentage of people failing the lecture on the first attempt between the two periods being compared), significant differences were found between the periods “Before CH” and “COVID-19” (p = 0.029) and between the periods “Before CH” and “After CH” (p < 0.001). The percentage of people failing the lecture on the first attempt in the period “Before CH” was significantly higher compared to the “COVID-19” and “After CH” periods. This percentage decreased from 7.36% (“Before CH”) to 3.42% (during the pandemic) to 1.68% (“After CH”).

Table 6. Assessment grades given in the first attempt by the teachers in the investigated periods.

		Grade
Teacher	Period	2.0	3.0	3.5	4.0	4.5	5.0	All
T1	Before CH	3	12	22	29	56	38	160
	COVID-19	11	9	9	48	25	46	148
	After CH	6	9	12	38	56	53	174
T2	Before CH	3	4	7	19	32	8	73
	COVID-19	2	6	11	22	26	50	117
	After CH	9	8	22	32	105	161	337
T3	Before CH	3	2	6	8	7	20	46
	After CH	-	-	-	8	9	45	62
T4	Before CH	4	5	19	14	35	11	88
	COVID-19	2	3	4	12	5	18	44
	After CH	1	5	1	8	17	24	56
T5	COVID-19	3	18	3	25	5	63	117
	After CH	1	3	1	9	15	57	86
T6	COVID-19	1	-	2	4	4	3	14

shows the assessment grades given in the first attempt by the teachers, divided by the investigated periods. Uniform grading guidelines and teaching materials, established for the teachers conducting the classes, minimized the bias resulting from the individual approach to conducting the classes. After examining the data contained in Table 6, it can be seen that for most teachers, the grades are distributed in a similar way. Teachers T3 and T5 stand out, with grade 5.0 dominating. These may be instructors’ subjective approach partially at fault (despite preventive measures), or the result of diversity lack within the particular laboratory groups. Each group is formed by approximately 15 students. Some groups may consist solely of more ambitious students, while others may be less committed (higher grades may prevail in better groups, lower grades in worse groups). Similarly, in the context of skills distribution among students. If these groups are assigned to different instructors, this may suggest increased grading bias towards the instructor.

The histograms presented in Figure 7 and Figure 8 illustrate the distribution of grades (assessment and exam, respectively) obtained in the first attempt in each of the analyzed periods. Both in the case of exam and assessment, grades 4.5 and 5.0 dominate (over 50% of all grades), with a noticeable increase in better grades (especially grade 5.0) after the introduction of cultural heritage elements to the curriculum in comparison to the period “Before CH”.

5. Conclusions and Discussion

Lublin University of Technology (Department of Computer Science, Faculty of Electrical Engineering and Computer Science, to be precise) has been offering computer science studies for over 25 years. From the beginning, computer graphics have played a crucial role in the studies program. The rapid development of 3D scanning techniques and Lab3D international efforts aimed at cultural heritage digitization have opened up new opportunities for enriching the regular introductory course of computer graphics with elements of digitized heritage. It was performed in a way that does not disturb the regular didactic process, nor the scope of the course learning outcomes. A research goal emerged, addressed in this article, to check what the effect is of such changes on students’ achievements.

We decided to analyze computer science students’ grades from 2018–2024; full-time studies were chosen as their source. In total, 3042 grades of 1522 students underwent the statistical analysis in order to support answering hypotheses. Six teachers were involved in the process. Analyzing just grades is not a perfect approach, although at the time of introducing the course modifications, nobody thought about a structured survey gathering subjective opinions of students; thus, these opinions are missing now. Another interesting factor is the COVID-19 pandemic, which was the reason behind dividing the analyzed years into three periods. Challenges and limitations faced during the curriculum implementation are already discussed in Section 3.5.

It was stated in H1 that the introduction of digitized cultural heritage elements to the computer graphics laboratory has a positive effect, which translates into better grades obtained by students at the end of laboratory classes. It was confirmed by the statistical test that a significant improvement in assessment grades (laboratory classes) occurred between the “Before CH” and “COVID-19” periods (in favor of the latter one), and next between “COVID-19” and “After CH” periods (again in favor of the latter one). The same result was obtained in the case of H2, stating that exam grades (lectures) obtained by students at the end of computer graphics lectures are better than before introducing digitized cultural heritage elements to the computer graphics course. The reason might be more successful obtainment of learning outcomes thanks to the more engaging, interesting course. Both H1 and H2 seem to be confirmed.

It was stated in H3 that the introduction of digitized cultural heritage elements to the computer graphics laboratory has a positive effect, which translates into better exam grades obtained by students at the end of lectures. The statistical analysis revealed a very strong positive and statistically significant correlation between assessment and exam grades. This means that students who received better assessment grades also received better exam grades. Again, it might be the result of more interesting classes, e.g., stimulating better memorability or encouraging students to dive into the theoretical aspects of computer graphics.

It was stated in H4 that the introduction of digitized cultural heritage elements to the computer graphics course has a positive effect, which translates into a significantly lower number of negative grades given to students during the first attempts to pass. It was confirmed by statistical tests that the number of negative exam grades was statistically significantly lower in “COVID-19” than in the “Before CH” period, and lower in the “After CH” period than in the “Before CH” period. The tests were inconclusive in the case of laboratory classes. Absolute numbers of negative assessment grades (13–19, depending on the period) and exam grades (12–27, depending on the period) were not too large. In the opinion of the authors, H4 is not convincingly confirmed.

It was stated in H5 that, due to the extraordinary circumstances of education, results obtained by students during the COVID-19 pandemic were different than after it finished. Statistical analysis of differences between “COVID-19” and “After CH” periods revealed that (a) exam grades were significantly better in the “After CH” period, (b) assessment grades were significantly better in the “After CH” period, (c) exam and assessment grades have significant very strong positive correlation, but the strength of correlation is marginally lower during the “COVID-19” period, and (d) the difference in number of negative grades (both assessment and exam grades) is insignificant. Statistical analysis of differences between “Before CH” and “COVID-19” periods revealed that (a) exam grades were significantly better in the “COVID-19” period, (b) assessment grades were significantly better in the “COVID-19” period, (c) exam and assessment grades have significant very strong positive correlation, but the strength of correlation is marginally higher during the “COVID-19” period, (d) number of negative assessment grades is insignificant, (e) number of negative exam grades is significantly lower in the “COVID-19” period. What is surprising is that bigger inconsistencies between the “COVID-19” and the other two periods were expected. Indeed, the “COVID-19” period was different (H5 is confirmed), although some general trends could be observed.

It was stated in H6 that introducing digitized cultural heritage elements to the computer graphics course makes it more interesting for the participating students. There are some general premises that may indicate confirmation of this hypothesis: (a) students obtained better grades because they were probably more engaged in the didactic process, (b) computer graphics teachers, who are co-authors of this article, noticed that more elaborate and complex student projects began to appear. Observations of the teachers were consistent—students’ interest in implementing the projects increased, the number and length of discussions (with a teacher and within a student group) on various solutions increased significantly, and the presented projects included animation solutions that had not been used by students before. Moreover, we assume by analogy to these works [45,46,47] that introducing real-life examples to the education process may increase the students’ engagement. Although we are inclined to confirm this hypothesis, its validation requires deeper research and broader analyses.

Unfortunately, currently structured surveys gathering students’ opinions are missing—it can only be imagined how the modified computer graphics course is perceived by them. The natural direction of future works is gathering such feedback, most likely at the end of the next course edition. Moreover, the authors decided that detailed class scenarios should not be included in the article for the purpose of maintaining its clarity. Trying to join these two areas in another work (article) sounds reasonable.

Nevertheless, the most important conclusion here is that introducing elements of digitized heritage did not worsen students’ achievements. The authors believe that the strength of their article is not only the proposal of the proven computer graphics curriculum, enriched with the digitized cultural heritage assets, but also the impact verification of introducing such content on students’ performance. The rarely encountered large research sample and long duration of observation cannot be underestimated either. Therefore, this article is of particular interest to persons who are searching for proven curriculum that can be adopted, or who want to make their curricula enriched with elements of digitized cultural heritage, but are afraid of the negative effects of such changes.