Intelligent and Soft Computing Systems for Green Energy

10Region-Based Convolutional Neural Networks for Selective Search

R. Kavitha¹*, Srinivasan R¹, P. Subha² and M. Kavitha¹

¹Vel Tech Rangarajan Dr. Sagunthala R & D Institute of Science and Technology, Chennai, Tamil Nadu, India

²Sri Sai Ram Institute of Technology, Chennai, India

Abstract

In recent years, image stitching and selective search using neural networks has had an increasingly significant role in various fields, including moving pictures, astronomy and healthcare. Image recognition through selective search consists of complex algorithms, and several cumbersome calculations produce “scans” which then merge together to form a real-life representation of the required area. This paper introduces a low-cost modeling method with user-friendly application that involves the concept of Image Stitching. It also discusses graphics rendering software with simulation of user movement in the scenario created on the computer. This study investigates a type of Harris picture stitching technique that is based on the OpenCV setup environment, in light of the immense scene and high-resolution image stitching challenges. To begin, the feature points are extracted using Harris corner detection. The feature points are then rough-matched using Normalized Cross Correlation, then the algorithm RANSAC is employed to eliminate incorrect matching. Second, to implement the image registration process, a cylindrical projection transformation model is used. Finally, to fuse photos, this study employs ...