Industry 4.0 transformation: adaptive coati deep convolutional neural network-based oral cancer diagnosis in histopathological images for clinical applications

Oral cancer is common cancer that appears in the mouth, posing a significant threat to public health due to its high mortality rate. Oral Squamous Cell Carcinoma (OSCC) is the most prevalent type of oral cancer, accounting for most cases, and it holds the seventh position among all types of cancers worldwide. Detecting OSCC early on is crucial to increase the chances of successful treatment and improve patients' survival rates. However, traditional diagnosis methods such as biopsy, where small tissue samples are extracted from the affected area and tested under a microscope, are time-consuming and require expert analysis. Moreover, due to the heterogeneity of OSCC, accurate diagnosis is challenging, and there is a need for alternative approaches to enhance the detection result of OSCC images. Therefore, this work develops two new approaches for segmenting and identifying OSCC with deep learning techniques named Mask Mean Shift CNN, named MMShift-CNN. The proposed MMShift-CNN approach attained the highest results in segmenting the OSCC region from the input image by retrieving color, texture, and shape features. The novel proposed method attained better performance with accuracy, F-measure, MSE, precision, sensitivity, and specificity of 0.9883, 0.9883, 0.0117, 0.999, 0.9867, and 0.99, respectively. These results reveal the efficiency of the proposed approach in accurately detecting oral cancer and potentially improving the efficiency of oral cancer diagnosis.