Deep Feature Learning Framework for Automated Oral Cancer Detection

This work presents a deep learning–based framework for the automatic identification of oral cancer from clinical tongue images. The study utilizes a multi-class oral image dataset comprising healthy tongue samples along with pathological conditions such as oral cancer, leukoplakia, oral lichen planus, thrush, and hairy tongue. Images collected from both affected and non-affected individuals were used to train and evaluate the proposed model. A pre-trained DenseNet169 network was adopted as the backbone architecture and fine-tuned using transfer learning, with additional fully connected layers introduced to enhance class discrimination. To reduce overfitting and improve generalization, extensive image augmentation techniques were applied during training. The effectiveness of the proposed approach was validated through a comparative analysis with a classical LeNet-based convolutional network. Experimental results indicate that the DenseNet-based model achieved superior performance, recording an accuracy of 94.08%, precision of 94.16%, recall of 94.70%, and an F1-score of 94.70%. In contrast, the LeNet model produced significantly lower results, with accuracy, precision, recall, and F1-score values close to 64%. The findings emphasize the importance of appropriate model architecture selection, robust data preprocessing, and systematic evaluation in medical image classification tasks. Further optimization and large-scale validation could strengthen the applicability of the proposed system for real-time clinical oral cancer screening.

Oral cancer detection, convolutional neural networks, medical image classification

1. P. F. Ndlovu, L. S. Magwaza, S. Z. Tesfay, and R. R. Mphahlele, “Rapid and non-destructive techniques for identifying adulteration in powdered horticultural products: A comprehensive review,” Food Research International, vol. 157, pp. 111198, 2022. [Google Scholar] [Crossref]

2. Y. Chen, Y. Wu, Y. Geng, X. Liu, L. Xiao, D. Xu, and H. Ma, “Non-invasive evaluation of anti-inflammatory activity and molecular indicators of oral solutions using near-infrared spectroscopy,” Molecules, vol. 27, no. 9, p. 2955, 2022. [Google Scholar] [Crossref]

3. M. M. Jadhav, A. O. Mulani, and M. Seth, “A painless, non-invasive machine learning approach for estimating blood glucose levels,” in IoT, Artificial Intelligence, and Smart Materials for Energy Applications, pp. 83–100, 2022. [Google Scholar] [Crossref]

4. Y. Qu, R. X. Xu, H. Fan, and Y. Meng, “Low-cost thermal imaging combined with machine learning for non-invasive pneumonia diagnosis and monitoring,” Infrared Physics & Technology, vol. 123, p. 104201, 2022. [Google Scholar] [Crossref]

5. S. X. Wang, D. L. Cortade, and A. S. de Olazarra, “Saliva-based non-invasive genotyping using magnetoresistive nanosensors and recombinase polymerase amplification,” Lab on a Chip, vol. 22, no. 11, pp. 2131–2144, 2022. [Google Scholar] [Crossref]

6. R. J. Price, V. R. Breza, K. M. Nowak, and M. R. Schwartz, “Recent developments in sonodynamic therapy for targeted, non-invasive cancer treatment,” Cancer Letters, vol. 532, p. 215592, 2022. [Google Scholar] [Crossref]

7. J. Yu, K. Shan, X. Shang, J. Guo, G. Liu, Z. Mu, and X. Liang, “Surface-enhanced Raman scattering for wearable and non-invasive medical diagnostics,” Frontiers in Chemistry, vol. 10, p. 1060322, 2022. [Google Scholar] [Crossref]

8. H. Sharghi, M. Javid, M. M. Bordbar, H. Samadinia, A. Sheini, J. Aboonajmi, and H. Bagheri, “A microfluidic colorimetric sensor array for non-invasive COVID-19 detection through urine analysis,” Microchimica Acta, vol. 189, no. 9, p. 315, 2022. [Google Scholar] [Crossref]

9. F. Sabbagh, I. I. Muhamad, R. Niazmand, P. K. Dikshit, and B. S. Kim, “Advances in polymer-based non-invasive insulin delivery systems,” International Journal of Biological Macromolecules, vol. 203, pp. 222–243, 2022. [Google Scholar] [Crossref]

10. C. Zambonin and A. Aresta, “Identification of cancer biomarkers from saliva and urine samples using MALDI-TOF/MS,” Molecules, vol. 27, no. 6, p. 1925, 2022. [Google Scholar] [Crossref]

11. S. W. Chang, S. Abdul-Kareem, A. F. Merican, and R. B. Zain, “Hybrid feature selection and machine learning techniques for oral cancer prognosis using clinicopathologic and genomic data,” BMC Bioinformatics, vol. 14, art. no. 170, May 2013. [Google Scholar] [Crossref]

12. H. Rajaguru and S. Prabhakar, “Comparative analysis of oral cancer classification using Gaussian mixture models and multilayer perceptrons,” in Proceedings of the International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2017, pp. 212–219. [Google Scholar] [Crossref]

• An Adaptive Joint Filtering Approach to Wireless Relay Network for Transmission Rate Maximization
• IoT-Integrated Mercury Substance Detection System for Cosmetic Product Safety
• Design and Implementation of Solar PV-Based Railway Microgrid for Linke Hofmann Busch Coaches
• Cost Control Techniques on Civil Engineering Projects in Oyo State, Nigeria
• Strength and Predictive Modeling of Corn Cob Ash Blended Concrete Using Multi-Output Artificial Neural Network Approach