PROJECT TITLE :

SDL Saliency-Based Dictionary Learning Framework for Image Similarity - 2018

ABSTRACT:

In image classification, getting adequate knowledge to learn a robust classifier has usually proven to be tough in many situations. Classification of histological tissue images for health care analysis could be a notable application in this context due to the requirement of surgery, biopsy or autopsy. To adequately exploit limited coaching information in classification, we propose a saliency guided dictionary learning technique and subsequently an image similarity technique for histo-pathological image classification. Salient object detection from pictures aids within the identification of discriminative image features. We tend to leverage the saliency values for the local image regions to learn a dictionary and respective sparse codes for an image, such that the more salient options are reconstructed with smaller error. The dictionary learned from an image offers a compact illustration of the image itself and is capable of representing pictures with similar content, with comparable sparse codes. We tend to use this concept to design a similarity measure between a pair of pictures, where local image features of 1 image, are encoded with the dictionary learned from the other and vice versa. To effectively utilize the learned dictionary, we take under consideration the contribution of each dictionary atom within the sparse codes to come up with a world image representation for image comparison. The efficacy of the proposed method was evaluated using three tissue data sets that consist of mammalian kidney, lung and spleen tissue, breast cancer, and colon cancer tissue pictures. From the experiments, we have a tendency to observe that our ways outperform the cutting-edge with a rise of 14.twopercent in the average classification accuracy over all knowledge sets.