Analysis of Histopathological Images for Early Diagnosis of Oral Squamous Cell Carcinoma by Hybrid Systems Based on CNN Fusion Features

Oral squamous cell carcinoma (OSCC) is one of the deadliest and most common types cancer. The incidence OSCC increasing annually, which requires early diagnosis to receive appropriate treatment. biopsy technique important techniques for analyzing samples, but it takes a long time get results. Manual still subject errors differences in doctors’ opinions, especially stages. Thus, automated can help doctors patients This study developed several hybrid models based on fused CNN features diagnosing OSCC-100x OSCC-400x datasets oral cancer, have ability analyze medical images with high level precision accuracy. They detect subtle patterns, abnormalities, or indicators diseases that may be difficult recognize naked eye. systems potential significantly reduce human error provide more consistent reliable results, resulting improved diagnostic also detection treatment success patient outcomes. By detecting at an stage, clinicians initiate interventions timely manner, potentially preventing progression improving chances successful first strategy was GoogLeNet, ResNet101, VGG16 pretrained, did not achieve satisfactory second adaptive region growing (ARG) segmentation algorithm. third mixed between ANN XGBoost networks ARG hashing fourth cancer by models. fusion GoogLeNet-ResNet101-VGG16 yielded AUC 98.85%, accuracy 99.3%, sensitivity 98.2%, 99.5%, specificity 98.35%.