Computer Vision

This project implements Zhang’s method for camera calibration. Camera calibration is essential in computer vision to obtain the intrinsic and extrinsic parameters of a camera, which are used to correct image distortions and understand the 3D geometry of the scene.

The panorama stitching project uses SIFT feature descriptors for key point matching and RANSAC for outlier rejection. The images are blended with distance transformation, resulting in a seamless and coherent panoramic image.

Convolution Neural Network architecture was designed to classify chats data into one of five categories (Vertical Bar, Horizontal Bar, Dot Line, Line, Pie). Fine tunned VGG16 model for this task. Implement GradCAM to visualize activation areas in images.