This repository contains a collection of Computer Vision projects. These projects cover a wide range of topics, including convolutional neural networks, image compressing, augmented reality, tracking and so on.
This was the most challenging but exciting project I have done in Computer Vision. The steps involved include calibrate the camera to obtain its intrinsic and extrinsic parameters, detect a target, and then, using a pose estimation algorithm, render a Pikachu over the target.
In this project, I used tracking algorithms to detect an object in the first frame of a video and follow it through subsequent frames.
Using convolutional neural networks, this project aims to automatically solve CAPTCHA challenges by recognizing and deciphering distorted text.
This project focuses on image compression, reducing file sizes without significantly compromising visual quality. To evaluate the compression, metrics such as compression rate, PSNR, and RMSE were used.
This is a conceptual project aimed at large-scale data collection using Computer Vision. The idea is based on the observation that button templates often repeat across various websites, allowing for the creation of a unique code, regardless of the HTML structure, to click in the button.