Computer Vision

Welcome to the Computer Vision repository! Here, you will find a variety of projects related to Computer Vision, an exciting field that combines image processing and machine learning to analyze, understand, and visually interpret the world around us.

1. Jupyter Notebooks

In this repository, you will find the following:

Getting-Started

• 1. Getting Started with Open CV.ipynb
• 2. Face Detection.ipynb
• 3 . Face Recognition.ipynb

Image-Classification

• Playing with Resnet-50.ipynb

2. Clasic Computer Vision

Preview	About
	💥 KNN with HOG 💥 This project implements a custom K-Nearest Neighbors (KNN) model for 10-class image classification using Histogram of Oriented Gradients (HOG) features. HOG captures edge structures, and KNN classifies images based on feature similarity, demonstrating a classic yet effective approach to image recognition. See details
	💥 Object tracking with Lucas Kanade 💥 This repository implements the Lucas-Kanade algorithm for object tracking, including standard, affine, and efficient inverse compositional variants, with mathematical explanations and visual results. See details
	💥 Homographies 💥 A homography is a transformation that maps points from one image to another by estimating a 3x3 projective transformation matrix. Given a set of corresponding points between two images, we solve for this matrix using techniques like Direct Linear Transformation (DLT) or RANSAC to handle outliers. Once computed, the homography allows us to warp one image to align with the other, enabling applications like image stitching, perspective correction, and augmented reality. See details

3. Neural Networks

Preview	About
	💥 Face Recognition 💥 Face recognition system. Search database engine plus similarity model to match faces. Additional face detection available for live video recognition integration system. See details
	💥 Image classification with CNNs 💥 This project serves as an introduction to PyTorch and Weights & Biases (wandb) by implementing and experimenting with deep learning models for image and text classification. The primary objectives include understanding PyTorch’s computation graphs, implementing a basic classifier, logging experiments with wandb, and modifying the baseline model to enhance performance. See details
	💥 Image generation with Diffusion models 💥 This project explores different generative models for image synthesis, including Convolutional Neural Networks (CNNs), Encoder-only Transformers, Generative Adversarial Networks (GANs), and Denoising Diffusion Probabilistic Models (DDPMs). See details