Design and Development of Content Based Image Retrieval Frameworks for
Remote Sensing Domain : Issues and Challenges

Dr. Tanish Zaveri, Professor, EC, IT Nirma University

To address the issues and challenges of CBIR system for remote sensing images, two
major frameworks are proposed. Framework-1 with the user defined feature sets and
Framework-2 using fast emerging Deep Learning Techniques. The dataset considered
for evaluating the proposed frameworks is UC Merced LULC data set, which is the
largest of its kind available in the public domain. Machine learning algorithms are
deployed to predict the appropriate distance metric. The evolving area of Deep
learning was deployed and experimented for CBIR system. Encouraging results were
published in the relevant journal.