MULTI-CHANNEL CONVOLUTION NEURAL NETWORK FOR ACCURATE CBMIR SYSTEM WITH REDUCED SEMANTIC GAP

The Content Based Medical Image Retrieval (CBMIR) system is an imaginative innovation to retrieve images from various types of medical images. The CBMIR is one of the applications of the Content Based Image retrieval (CBIR). To shrink the semantic gap in the domain of medical image for the precise diagnoses of the disease from historical cases is the major challenge in the CBMIR system. The semantic gap reduction is mainly depending on the feature representation of the images. The MultiChannel Convolution Neural Network (MC-CNN) model is utilized to get multiple features from the training images. The fusion of these features is being used for the training of the network. The trained model is used for the retrieval of the medical images and its relevant treatment, for the provided testing image of the patient. It helps the doctor to give proper treatment to the patient based on the past success ratio of that particular treatment. The average precision rate is the measurement parameter which is utilized to compare the proposed model with various transfer learning models. The proposed model is evaluated for three different medical datasets and it shows noteworthy improvement over various transfer learning models