Downloads: 1 | Views: 26 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Doctoral Thesis | Computer Science and Information Technology | India | Volume 13 Issue 4, April 2024
Classification of Oral Lesions using Deep Learning: A Retrospective Study
Hemashree H C | Pradeep N [2]
Abstract: The sixth most frequent cancer with a high death rate from disease - related causes is oral cancer. The main cause is that almost two thirds of patients receive a cancer diagnosis later in the disease's progression. Oral potentially malignant disorders (OPMDs) are observable alterations in the oral mucosa that occur before most oral malignancies. In 88% of cases, early detection of OPMDs can prevent the development of malignancy. The use of artificial intelligence (AI) in medicine, particularly oncology, has grown in popularity since it has shown useful in diagnosing and predicting the prognosis of cancer. This study uses a dataset of clinically annotated photographic pictures to identify oral precancerous and cancerous lesions and distinguish them from normal mucosa using pre - trained convolutional neural networks (CNNs). Clinical images of patients with oral squamous cell carcinoma and OPMDs were used in this investigation. Photographs of the normal oral mucosa were compared with these images for analysis. Image categorization was done by transfer learning with different CNN architectures that had already been trained. Regarding performance against other models, VGG19 performed well in the current investigation.
Keywords: Oral cancer, Artificial intelligence, Diagnosis, Image classification, Cancer prognosis
Edition: Volume 13 Issue 4, April 2024,
Pages: 185 - 193