Scholarlink

Oral cancer ranks as one of the most common cancers across the globe, holding the sixth position worldwide and the third in India. Squamous Cell Carcinoma represents about 90% of cases of oral cancer. Although the overall five-year survival rate is approximately 60%, timely detection can greatly enhance survival rates to exceed 90%. Traditional diagnostic methods, typically reliant on pathologists’ visual evaluations, can lead to delays and varied interpretations. This study presents an innovative deep learning framework that integrates Convolutional Neural Networks (CNNs) with Long Short-Term Memory (LSTM) networks for the automatic detection of oral cancer through microscopic images. The model employs EfficientNetB3 to extract image features and uses LSTM layers for analyzing temporal patterns, resulting in enhanced classification performance. By utilizing the Adam algorithm for optimization and categorical cross-entropy loss, the model reached an impressive training accuracy of 98.26%. Future directions involve broadening the dataset and incorporating the model into clinical processes to improve early and precise diagnostic abilities.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Oral Cancer Prediction Using Hybridize Neural Network Model: CNN and RNN

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ganeshayya Shidaganti,
  • Akshatha Kamath,
  • Vinod Sajjan,
  • Arjun M.,
  • Mohana N.,
  • Thejas

摘要

Oral cancer ranks as one of the most common cancers across the globe, holding the sixth position worldwide and the third in India. Squamous Cell Carcinoma represents about 90% of cases of oral cancer. Although the overall five-year survival rate is approximately 60%, timely detection can greatly enhance survival rates to exceed 90%. Traditional diagnostic methods, typically reliant on pathologists’ visual evaluations, can lead to delays and varied interpretations. This study presents an innovative deep learning framework that integrates Convolutional Neural Networks (CNNs) with Long Short-Term Memory (LSTM) networks for the automatic detection of oral cancer through microscopic images. The model employs EfficientNetB3 to extract image features and uses LSTM layers for analyzing temporal patterns, resulting in enhanced classification performance. By utilizing the Adam algorithm for optimization and categorical cross-entropy loss, the model reached an impressive training accuracy of 98.26%. Future directions involve broadening the dataset and incorporating the model into clinical processes to improve early and precise diagnostic abilities.