Melanoma Skin Cancer Recognition with a Convolutional Neural Network and Feature Dimensions Reduction with Aquila Optimizer
Abstract
:1. Introduction
- Use the CNN for feature extraction to avoid noisy features;
- Use the accurate features by using the metaheuristic method;
- Obtain the high level of accuracy required for diagnosis.
1.1. Contribution
- Novel Approach for Melanoma Recognition: In the present study, a new approach is proposed for recognizing melanoma by combining the CNN-based feature extraction method with the AO. The approach will focus on reducing dimensionality for improved effectiveness and efficiency in melanoma detection.
- CNN and AO integration: A system that takes the strengths of CNN in auto-feature extraction and those of AO in reducing feature dimensionality. The result of this combination is usually a compact and informative feature space, leading to improved system performance. It may also be suitable for running in resource-constrained environments.
- Thorough Experimental Evaluation: The proposed approach is evaluated very well on benchmark datasets, wherein its performance is measured regarding accuracy, sensitivity, and specificity. The results are compared against existing state-of-the-art methods to prove the efficacy of the proposed framework.
- Addressing Limitations of Traditional Methods: The present study performs a standard feature extraction to improve on the limitations of traditional methods, namely entropy, energy, or momentum, by the use of CNNs for the proper elimination of blurred noise to increase the reliability of melanoma classification.
1.2. Originality
2. Literature Review
3. Material and Methods
3.1. Dataset
3.2. Preprocessing
3.3. Convolutional Neural Network (CNN) Architecture
3.4. Feature Dimension Reduction with Aquila Optimizer
3.5. Performance Evaluation Metrics
3.6. Ablation Study for AO Efficiency
3.7. Evaluation Metrics
3.8. Cross-Validation
3.9. Implementation
4. Findings and Analysis
- Error Bars for Performance Metrics:
- 2.
- Training and Test Trend Analysis:
- 3.
- Confusion Matrix for Classification Quality:
5. Discussion on Model Limitations
- Dataset Bias: The datasets employed, including ISIC 2019, ISBI 2016, ISBI 2017, and PH2, do not cover the spectrum of melanoma types or other skin lesions. They primarily focus on specific lesion categories, which could introduce bias if the training data do not adequately represent the diversity found in real-word cases. Future studies could consider incorporating additional datasets to capture a wider range of skin lesions.
- Risk of Overfitting: Despite the dimensionality reduction achieved with the AO, the model may still be vulnerable to overfitting, particularly if the dataset is limited or the features are too complex. While the current methodology mitigates overfitting through data augmentation and careful feature selection, further validation on an external dataset is necessary to confirm the model’s generalization ability.
- Computational Demands: The use of DL techniques such as CNN along with metaheuristic techniques such as AO requires significant computational resources. While reducing feature dimensionality helps improve efficiency, training large models may still pose challenges in resource-constrained environments, such as mobile or edge computing platforms. Further research could focus on optimizing these methods to reduce computational requirements without compromising classification performance.
- Algorithmic Constraints: The performance of the AO algorithm is sensitive to its initial configuration and optimization process. While the method has shown significant improvement, it may not always yield the best results in every scenario. Further refinement or the adoption of hybrid strategies could improve its robustness and applicability across a wider range of datasets.
6. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Babic, Z.; Rancic, N.; Ilic, S.; Corovic, S.; Milicic, V.; Djordjevic, G.; Djordjevic, O.; Ilic, S.; Radevic, S. Trends in Melanoma Mortality in Serbia: A 22-Year Population-Based Study. Iran. J. Public Health 2024, 53, 828–836. [Google Scholar] [CrossRef] [PubMed]
- Nanz, L.; Keim, U.; Katalinic, A.; Meyer, T.; Garbe, C.; Leiter, U. Epidemiology of keratinocyte skin cancer with a focus on cutaneous squamous cell carcinoma. Cancers 2024, 16, 606. [Google Scholar] [CrossRef]
- Rohtash, R. Integrating Artificial Intelligence and Neuroscience: Advancements in Cognitive Computing. J. Asian Res. Found. 2025, 1, 1–22. [Google Scholar]
- Jaison, A.; Mohan, A.; Lee, Y.-C. Machine learning-enhanced photocatalysis for environmental sustainability: Integration and applications. Mater. Sci. Eng. R Rep. 2024, 161, 100880. [Google Scholar] [CrossRef]
- Binetti, M.S.; Massarelli, C.; Uricchio, V.F. Machine Learning in Geosciences: A Review of Complex Environmental Monitoring Applications. Mach. Learn. Knowl. Extr. 2024, 6, 1263–1280. [Google Scholar] [CrossRef]
- Khan, F.S.; Masum, A.A.; Adam, J.; Karim, M.R.; Afrin, S. AI in Healthcare Supply Chain Management: Enhancing Efficiency and Reducing Costs with Predictive Analytics. J. Comput. Sci. Technol. Stud. 2024, 6, 85–93. [Google Scholar] [CrossRef]
- Oliva, D.; Cuevas, E.; Pajares, G.; Zaldivar, D.; Osuna, V. A multilevel thresholding algorithm using electromagnetism optimization. Neurocomputing 2014, 139, 357–381. [Google Scholar] [CrossRef]
- Mesejo, P.; Ibáñez, O.; Cordón, O.; Cagnoni, S. A survey on image segmentation using metaheuristic-based deformable models: State of the art and critical analysis. Appl. Soft Comput. 2016, 44, 1–29. [Google Scholar] [CrossRef]
- Tajvidi Asr, R.; Rahimi, M.; Hossein Pourasad, M.; Zayer, S.; Momenzadeh, M.; Ghaderzadeh, M. Hematology and Hematopathology Insights Powered by Machine Learning: Shaping the Future of Blood Disorder Management. Iran. J. Blood Cancer 2024, 16, 9–19. [Google Scholar] [CrossRef]
- Dobre, E.-G.; Surcel, M.; Constantin, C.; Ilie, M.A.; Caruntu, A.; Caruntu, C.; Neagu, M. Skin cancer pathobiology at a glance: A focus on imaging techniques and their potential for improved diagnosis and surveillance in clinical cohorts. Int. J. Mol. Sci. 2023, 24, 1079. [Google Scholar] [CrossRef]
- Vayadande, K.; Bhosle, A.A.; Pawar, R.G.; Joshi, D.J.; Bailke, P.A.; Lohade, O. Innovative approaches for skin disease identification in machine learning: A comprehensive study. Oral Oncol. Rep. 2024, 10, 100365. [Google Scholar] [CrossRef]
- Naeem, A.; Anees, T.; Khalil, M.; Zahra, K.; Naqvi, R.A.; Lee, S.-W. SNC_Net: Skin Cancer Detection by Integrating Handcrafted and Deep Learning-Based Features Using Dermoscopy Images. Mathematics 2024, 12, 1030. [Google Scholar] [CrossRef]
- Attallah, O. Skin-CAD: Explainable deep learning classification of skin cancer from dermoscopic images by feature selection of dual high-level CNNs features and transfer learning. Comput. Biol. Med. 2024, 178, 108798. [Google Scholar] [CrossRef] [PubMed]
- Saleh, N.; Hassan, M.A.; Salaheldin, A.M. Skin cancer classification based on an optimized convolutional neural network and multicriteria decision-making. Sci. Rep. 2024, 14, 17323. [Google Scholar] [CrossRef] [PubMed]
- Saha, S.; Ghosh, M.; Ghosh, S.; Sen, S.; Singh, P.K.; Geem, Z.W.; Sarkar, R. Feature selection for facial emotion recognition using cosine similarity-based harmony search algorithm. Appl. Sci. 2020, 10, 2816. [Google Scholar] [CrossRef]
- Ahmadirad, Z. The Role of AI and Machine Learning in Supply Chain Optimization. Int. J. Mod. Achiev. Sci. Eng. Technol. 2025, 2, 1–8. [Google Scholar]
- Karaboga, D. An Idea Based on Honey Bee Swarm for Numerical Optimization; Technical Report—TR06; Department of Computer Engineering, Engineering Faculty, Erciyes University: Kayseri, Turkey, 2005. [Google Scholar]
- Zhu, G.; Kwong, S. Gbest-guided artificial bee colony algorithm for numerical function optimization. Appl. Math. Comput. 2010, 217, 3166–3173. [Google Scholar] [CrossRef]
- Al-Masni, M.A.; Al-Antari, M.A.; Choi, M.-T.; Han, S.-M.; Kim, T.-S. Skin lesion segmentation in dermoscopy images via deep full resolution convolutional networks. Comput. Methods Programs Biomed. 2018, 162, 221–231. [Google Scholar] [CrossRef]
- Esteva, A.; Kuprel, B.; Novoa, R.A.; Ko, J.; Swetter, S.M.; Blau, H.M.; Thrun, S. Dermatologist-level classification of skin cancer with deep neural networks. Nature 2017, 542, 115–118. [Google Scholar] [CrossRef]
- Hammouche, K.; Diaf, M.; Siarry, P. A multilevel automatic thresholding method based on a genetic algorithm for a fast image segmentation. Comput. Vis. Image Underst. 2008, 109, 163–175. [Google Scholar] [CrossRef]
- Sun, G.; Zhang, A.; Yao, Y.; Wang, Z. A novel hybrid algorithm of gravitational search algorithm with genetic algorithm for multi-level thresholding. Appl. Soft Comput. 2016, 46, 703–730. [Google Scholar] [CrossRef]
- Gao, H.; Xu, W.; Sun, J.; Tang, Y. Multilevel thresholding for image segmentation through an improved quantum-behaved particle swarm algorithm. IEEE Trans. Instrum. Meas. 2009, 59, 934–946. [Google Scholar] [CrossRef]
- Liu, Y.; Mu, C.; Kou, W.; Liu, J. Modified particle swarm optimization-based multilevel thresholding for image segmentation. Soft Comput. 2015, 19, 1311–1327. [Google Scholar] [CrossRef]
- Castillo, O.; Neyoy, H.; Soria, J.; Melin, P.; Valdez, F. A new approach for dynamic fuzzy logic parameter tuning in ant colony optimization and its application in fuzzy control of a mobile robot. Appl. Soft Comput. 2015, 28, 150–159. [Google Scholar] [CrossRef]
- Taherdangkoo, M.; Bagheri, M.H.; Yazdi, M.; Andriole, K.P. An effective method for segmentation of MR brain images using the ant colony optimization algorithm. J. Digit. Imaging 2013, 26, 1116–1123. [Google Scholar] [CrossRef] [PubMed]
- Horng, M.-H. Multilevel thresholding selection based on the artificial bee colony algorithm for image segmentation. Expert Syst. Appl. 2011, 38, 13785–13791. [Google Scholar] [CrossRef]
- Jiang, Y.; Hao, Z.; Yang, Z.; Wang, Y.; He, H. A cooperative honey bee mating algorithm and its application in multi-threshold image segmentation. In Proceedings of the 2014 IEEE Congress on Evolutionary Computation (CEC), Beijing, China, 6–11 July 2014; pp. 1579–1585. [Google Scholar]
- Yang, X.-S. A new metaheuristic bat-inspired algorithm. In Nature Inspired Cooperative Strategies for Optimization (NICSO 2010); Springer: Berlin/Heidelberg, Germany, 2010; pp. 65–74. [Google Scholar]
- Ye, Z.-W.; Wang, M.-W.; Liu, W.; Chen, S.-B. Fuzzy entropy based optimal thresholding using bat algorithm. Appl. Soft Comput. 2015, 31, 381–395. [Google Scholar] [CrossRef]
- Maitra, M.; Chatterjee, A. A novel technique for multilevel optimal magnetic resonance brain image thresholding using bacterial foraging. Measurement 2008, 41, 1124–1134. [Google Scholar] [CrossRef]
- Yang, C.; Ji, J.; Liu, J.; Liu, J.; Yin, B. Structural learning of Bayesian networks by bacterial foraging optimization. Int. J. Approx. Reason. 2016, 69, 147–167. [Google Scholar] [CrossRef]
- Fister, I.; Fister Jr, I.; Yang, X.-S.; Brest, J. A comprehensive review of firefly algorithms. Swarm Evol. Comput. 2013, 13, 34–46. [Google Scholar] [CrossRef]
- Chen, S.; Yao, L.; Chen, B. A parameterized logarithmic image processing method with Laplacian of Gaussian filtering for lung nodule enhancement in chest radiographs. Med. Biol. Eng. Comput. 2016, 54, 1793–1806. [Google Scholar] [CrossRef] [PubMed]
- Cheng, L.; Wu, X.; Wang, Y. Artificial flora (AF) optimization algorithm. Appl. Sci. 2018, 8, 329. [Google Scholar] [CrossRef]
- Zhang, J.; Li, H.; Tang, Z.; Lu, Q.; Zheng, X.; Zhou, J. An Improved Quantum-Inspired Genetic Algorithm for Image Multilevel Thresholding Segmentation. Math. Probl. Eng. 2014, 2014, 295402. [Google Scholar] [CrossRef]
- Olugbara, O.O.; Ndhlovu, B.N. Constructing frugal sales system for small enterprises. Afr. J. Inf. Syst. 2014, 6, 1. [Google Scholar]
- Mendonça, T.; Ferreira, P.M.; Marques, J.S.; Marcal, A.R.S.; Rozeira, J. PH2—A dermoscopic image database for research and benchmarking. In Proceedings of the 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Osaka, Japan, 3–7 July 2013; pp. 5437–5440. [Google Scholar]
- Abualigah, L.; Yousri, D.; Abd Elaziz, M.; Ewees, A.A.; Al-Qaness, M.A.A.; Gandomi, A.H. Aquila optimizer: A novel meta-heuristic optimization algorithm. Comput. Ind. Eng. 2021, 157, 107250. [Google Scholar] [CrossRef]
- Menegola, A.; Fornaciali, M.; Pires, R.; Bittencourt, F.V.; Avila, S.; Valle, E. Knowledge transfer for melanoma screening with deep learning. In Proceedings of the 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017), Melbourne, VIC, Australia, 18–21 April 2017; pp. 297–300. [Google Scholar]
- Barata, C.; Celebi, M.E.; Marques, J.S. Improving dermoscopy image classification using color constancy. IEEE J. Biomed. Health Inform. 2014, 19, 1146–1152. [Google Scholar]
- Xie, F.; Fan, H.; Li, Y.; Jiang, Z.; Meng, R.; Bovik, A. Melanoma classification on dermoscopy images using a neural network ensemble model. IEEE Trans. Med. Imaging 2016, 36, 849–858. [Google Scholar] [CrossRef]
- Vasconcelos, C.N.; Vasconcelos, B.N. Experiments using deep learning for dermoscopy image analysis. Pattern Recognit. Lett. 2020, 139, 95–103. [Google Scholar] [CrossRef]
- Oliveira, R.B.; Pereira, A.S.; Tavares, J.M.R.S. Computational diagnosis of skin lesions from dermoscopic images using combined features. Neural Comput. Appl. 2019, 31, 6091–6111. [Google Scholar] [CrossRef]
- Guo, Y.; Ashour, A.S.; Smarandache, F. A novel skin lesion detection approach using neutrosophic clustering and adaptive region growing in dermoscopy images. Symmetry 2018, 10, 119. [Google Scholar] [CrossRef]
- Li, Y.; Shen, L. Skin lesion analysis towards melanoma detection using deep learning network. Sensors 2018, 18, 556. [Google Scholar] [CrossRef]
- Bi, L.; Kim, J.; Ahn, E.; Feng, D. Automatic skin lesion analysis using large-scale dermoscopy images and deep residual networks. arXiv 2017, arXiv:1703.04197. [Google Scholar]
References | Algorithm | Method |
---|---|---|
[17,18,19,20] | Artificial Bee Colony (ABC) | Inspired by the intelligent behavior of honeybees |
[21,22] | Genetic Algorithm (GA) | Imitates the process of natural selection |
[23,24] | Particle Swarm Optimization (PSO) | Based on the social behavior of fish schools and bird flocks |
[25,26] | Ant Colony Optimization (ACO) | Choosing a route from its nest to a food supply is based on how ants forage. |
[27,28] | Honey bee mating optimization algorithm (HBMO) | Inspired by the process of mating in real honey bees. |
[29,30] | Bat Algorithm (BA) | Influenced by how microbats use echolocation |
[31,32] | Bacterial colony optimization (BCO) | Uses the whole life cycle of the E. coli bacteria to simulate some of their normal behavior. |
[33,34] | Firefly Algorithm (FA) | Motivated by the tropical firefly’s flashing light displays |
[35] | Artificial Flora (AF) | Takes the Artificial Flora procedures as an inspiration. |
Dataset | Number of Classes | Samples per Class |
---|---|---|
ISIC 2019 | 2 | 1000 benign, 1000 malignant |
ISBI 2016 | 2 | 200 benign, 200 malignant |
ISBI 2017 | 2 | 300 benign, 300 malignant |
PH2 Dataset | 3 | 35 melanocytic nevi, 25 dysplastic nevi, 30 melanomas |
Dataset | Features Before AO | Features After AO | Reduction Percentage |
---|---|---|---|
ISIC 2019 | 1024 | 256 | 75% |
ISBI 2016 | 1024 | 300 | 70.70% |
ISBI 2017 | 1024 | 280 | 72.70% |
Dataset | Accuracy | Sensitivity | Specificity | Precision | F1 Score | AUC |
---|---|---|---|---|---|---|
ISIC 2019 | 92.50% | 93.20% | 91.80% | 92.00% | 92.60% | 0.96 |
ISBI 2016 | 90.00% | 91.50% | 88.60% | 90.40% | 90.90% | 0.94 |
ISBI 2017 | 91.20% | 92.10% | 89.40% | 91.10% | 91.50% | 0.95 |
PH2 Dataset | 93.00% | 94.40% | 91.90% | 93.50% | 93.70% | 0.97 |
Dataset | Accuracy (w/o AO) | Accuracy (w/AO) | Sensitivity (w/o AO) | Sensitivity (w/AO) | Specificity (w/o AO) | Specificity (w/AO) | AUC (w/o AO) | AUC (w/AO) |
---|---|---|---|---|---|---|---|---|
ISIC 2019 | 88.40% | 92.50% | 89.10% | 93.20% | 86.80% | 91.80% | 0.91 | 0.96 |
ISBI 2016 | 85.70% | 90.00% | 87.30% | 91.50% | 84.10% | 88.60% | 0.89 | 0.94 |
ISBI 2017 | 87.20% | 91.20% | 88.00% | 92.10% | 85.50% | 89.40% | 0.9 | 0.95 |
PH2 Dataset | 89.10% | 93.00% | 90.50% | 94.40% | 87.80% | 91.90% | 0.92 | 0.97 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Mohamed, J.; Tezel, N.S.; Rahebi, J.; Ghadami, R. Melanoma Skin Cancer Recognition with a Convolutional Neural Network and Feature Dimensions Reduction with Aquila Optimizer. Diagnostics 2025, 15, 761. https://doi.org/10.3390/diagnostics15060761
Mohamed J, Tezel NS, Rahebi J, Ghadami R. Melanoma Skin Cancer Recognition with a Convolutional Neural Network and Feature Dimensions Reduction with Aquila Optimizer. Diagnostics. 2025; 15(6):761. https://doi.org/10.3390/diagnostics15060761
Chicago/Turabian StyleMohamed, Jalaleddin, Necmi Serkan Tezel, Javad Rahebi, and Raheleh Ghadami. 2025. "Melanoma Skin Cancer Recognition with a Convolutional Neural Network and Feature Dimensions Reduction with Aquila Optimizer" Diagnostics 15, no. 6: 761. https://doi.org/10.3390/diagnostics15060761
APA StyleMohamed, J., Tezel, N. S., Rahebi, J., & Ghadami, R. (2025). Melanoma Skin Cancer Recognition with a Convolutional Neural Network and Feature Dimensions Reduction with Aquila Optimizer. Diagnostics, 15(6), 761. https://doi.org/10.3390/diagnostics15060761