Categorization of Digital Pathology Image using Deep Learning model

Computer-aided detection and diagnosis have transformed the medical research landscape, particularly in colorectal carcinoma. This paper introduces a novel approach to address the multi-class challenge in colorectal cancer detection using a fully convolutional network augmented with a pre-trained convolutional neural network and a spatial attention module. Specifically, the proposed model employs ResNet-50 in the contracting path of the fully convolutional network, leading to dimensionality reduction and feature extraction. A unique dimensionality reduction technique utilizing standard deviation is introduced to optimize data representation while managing computational complexity. The model’s architectural design involves a carefully curated head part with fourteen layers, including a dropout layer and dense layers with rectified linear units and softmax activation functions for accurate image categorization. The proposed approach demonstrates promising results in addressing the challenges associated with colorectal cancer detection in digital pathology, showcasing the potential of deep learning and attention mechanisms in enhancing classification accuracy.

• colorectal cancer detection
• digital pathology images
• computer-aided diagnosis
• multi-class image classification
• fully convolutional network
• spatial attention module
• ResNet50
• dimensionality reduction
• attention mechanism

• 3 — Bonne santé et bien-être

• AI in cancer detection
• Advanced Neural Network Applications
• Colorectal Cancer Screening and Detection
• colorectal cancer detection
• digital pathology images
• computer-aided diagnosis
• multi-class image classification
• fully convolutional network
• spatial attention module
• ResNet50
• dimensionality reduction
• Artificial Intelligence
• Computer Vision and Pattern Recognition
• Oncology
• Computer Science
• Medicine