1. Introduction
Among other cancers, skin cancer is considered one of the deadliest diseases. Around 1.2 million people died in 2020 due to skin cancer only [
1]. According to the WHO [
1], skin cancer was one of the most common cancers in terms of new cases in 2020, and the number of new cases is increasing dramatically [
2,
3]. One of the common causes of skin cancer is exposure of the skin to UV (ultraviolet) rays directly coming from the sun [
4]. It is said that such rays affect fair-skinned people and those with sensitive skin more than dark-skinned ones [
5].
Most deaths are caused by invasive melanoma, which constitutes only 1% of total skin cancer cases. From historical data, it is found that melanoma skin cancer cases are rising rapidly. According to the most recent report from the American Cancer Society, which provides data up until 2022 [
6], it was estimated that approximately 99,780 cases of melanoma cancer would have been diagnosed by the end of the year, with 57,180 cases among men and 42,600 cases among women. The report also indicated that around 7650 deaths were expected due to melanoma cancer, with approximately 5080 deaths among men and 2570 deaths among women.
To cure any cancer, it is best to detect it at an early stage, and skin cancer is no different. Any unusual growth or new/changing skin spots must be evaluated. If there are any new lesions or any change in a lesion’s appearance, whether in size, color, or shape, it should be shown to a doctor and evaluated accordingly. To detect skin cancer, doctors use multiple techniques, and one of the ways is visual detection [
7]. A manual has been developed by the American Center for the Study of Dermatology, and is used by doctors for initial screening. This manual is called asymmetry, border, color, and diameter (ABCD). At the initial stage, the doctor suspects a skin lesion on the patient’s body and recommends going for a biopsy [
8]. The reports are examined, and a thorough check is performed to detect whether it is benign or malignant and the type of cancer [
9]. Another technique, called dermoscopy, can be used to diagnose skin cancers [
10]. In this technique, bright images of the skin lesion are captured, highlighting dark spots [
11]. Nevertheless, these methods are inefficient because they cannot help diagnose the nature of the lesion. This is due to many reasons such as the presence of blood vessels or hair around the lesion, the intensity of the light, failure to correctly capture the shape of the lesion, and not identifying cancerous and non-cancerous lesions correctly [
12,
13,
14].
The average accuracy of diagnosing skin cancers by manually examining dermoscopic images is 60% to 80%. The accuracy varies from one dermatologist to another based on their years of experience. It has been claimed that a dermatologist with three to five years of experience can have an accuracy of around 60%, whereas there is an improvement in accuracy by 20% if the dermatologist has 10+ years of experience [
15]. Therefore, it can be claimed that dermoscopy requires extensive training to yield better diagnosis. Different types of skin cancers can be identified with the help of dermoscopic images. However, there are two main types of skin cancers: melanocytic and nonmelanocytic. Melanotic skin cancers consist only of melanoma and melanocytic nevi. However, nonmelanocytic skin cancers have many types, such as dermatofibroma (DF), vascular (VASC), benign keratosis lesions (BKL), basal cell carcinoma (BCC), and squamous cell carcinoma (SCC) [
16].
Melanoma is a type of skin cancer arising from abnormal melanin production in melanocyte cells. It is the most prevalent and lethal form of skin cancer and is categorized into benign and malignant types [
17]. While benign melanoma lesions contain melanin in the epidermal layer, malignant melanomas display excessive melanin production. The United States reports over five million new cases of skin cancer each year, with melanoma accounting for three-quarters of all skin cancer fatalities, resulting in 10,000 deaths annually [
18]. In 2021, the US registered 106,110 cases of melanoma, leading to 7180 fatalities, with projections indicating a 6.5% increase in melanoma-caused deaths in 2022. In 2022, it is expected that 197,700 new cases of melanoma will be diagnosed in the US alone [
19]. Every year, around 100,000 new cases of melanoma are discovered throughout Europe [
20]. Melanoma is detected in 15,229 people in Australia each year [
18,
21]. Skin cancer incidence rates have climbed in the last decade, with melanoma rates increasing by 255% in the United States and 120% in the United Kingdom since the 1990s [
22,
23]. Melanoma, however, is considered a highly curable cancer if detected early. In the early stages, survival rates exceed 96%. In the advanced stage, by contrast, survival rates drop to 5%. When melanoma has spread throughout the body, treatment becomes more difficult [
16].
The adoption of artificial intelligence (AI) and deep learning [
24,
25] has resulted in significant advancements in the accuracy and efficiency of skin cancer classification, assisting in the disease’s early diagnosis and treatment [
26]. When trained on massive datasets of skin scans, AI systems may learn to recognize the characteristics of malignant cells and distinguish them from benign cells with high accuracy. Several studies have explored using AI and deep learning for skin cancer classification [
27,
28]. Khan et al. [
29] adopted the DarkeNet19 model and trained it on multiple datasets such as HAM10000, ISBI2018, and ISBI2019. They fine-tuned this model and achieved 95.8%, 97.1%, and 85.35% accuracy for the HAM10000, ISBI2018, and ISBI2019 datasets, respectively. In comparison, another study [
30] trained three different models (InceptionV3, ResNet, and VGG19) on a dataset containing 24,000 images retrieved between 2019 and 2020 from the ISIC archive. They concluded that InceptionV3 outperformed the rest of the models regarding accuracy. On the other hand, Khamparia et al. [
31] incorporated transfer learning while training different deep learning architectures and proved that transfer learning and data augmentation helped to improve the results.
When training deep learning models, data imbalance is always an issue. There are many ways by which authors improve datasets by incorporating data augmentation techniques [
32]. Ahmad et al. [
33] used a data augmentation technique called generative adversarial networks (GAN), which creates artificial images similar to the original images to improve the dataset. With the help of this technique, they claim that their model accuracy was enhanced from 66% to 92%. In another study Kausar et al. [
34] used some fine-tuning techniques to improve state-of-art deep learning image classification models. They achieved an accuracy of 72%, 91%, 91.4%, 91.7%, and 91.8% for ResNet, InceptionV3, DenseNet, InceptionResNetV2, and VGG-19, respectively. Khan et al. [
35] proposed a multiclass deep learning model trained on the HAM10000, ISBI2018, and ISIC2019 datasets. They also incorporated transfer learning, and their results showed that the proposed model achieved an accuracy of 96.5%, 98%, and 89% for the HAM10000, ISBI2018, and ISIC2019 datasets, respectively. In another study, Deepa et al. [
36] trained the ResNet50 model on the International Skin Image Collaboration (ISIC) dataset and achieved 89% accuracy. Tahir et al. [
37] proposed a deep learning model called DSCC_Net, trained that on three datasets, ISIC 2020, HAM10000, and DermIS, and achieved an accuracy of 99%. They further compared their model with other state-of-art models and concluded that it outperformed them all. In another study, Shaheen et al. [
38] proposed a multiclass model using particle swarm optimization trained on the HAM1000 dataset. They claim that their model achieved 97.82% accuracy.
As described above, there is a general trend for image processing based on deep learning to gradually adopt deeper networks. The benefit of using deeper networks is obvious, i.e., a deeper network provides stronger nonlinear representation capability. This means that, for some specific problems, a deeper network may be better able to learn more complex transformations and thus fit more complex feature inputs. However, previous research (see, e.g., [
39]) has also shown ways in which network depth may negatively affect classification performance in cases where relatively simpler features are involved. Here, we first quantitatively assess the effect of network depth on classification performance and then develop a shorter and broader variant of the originally selected model (termed SBXception). The main contributions of this paper are the following:
We analyze the characteristics of the adopted dataset (HAM10000) to show that network depth beyond an optimal level may not be suitable for classification tasks on this dataset.
A new, shorter, broader variant of the Xception model is proposed to classify various skin lesions efficiently.
The proposed modified model architecture is used to provide better classification performance compared to the state-of-the-art methods.