This repo contains the collection of projects designed and executed during my term with Udacity's Computer vision nano degree program
- Project1 - Detecting facial keypoints
- Facial keypoints define certain landmarks on the face such as eyes, nose, mouth, etc. These are usually usefull information in applications such as face detection, emotion recognition, expression recognition, etc.
- Project2 - Image Captioning
- Image captioning is an useful service that helps in describing photos automatically on social media or to visually disabled people. This project uses an encoder-decoder architecture to implement a simple caption generator.
- Project3 - Landmark Detection & Tracking
- Object localization & tracking is one of the biggest topic of research in the present day. SLAM is used mainly in robotics & autonomous driving vehicles. This project explores the implimentation of SLAM technique for locating landmarks around the object of interest and determining its location post sensor readings.
- Pytorch Practice
- The framework used in the nano degree was pytorch. A hands-on semi-course was provided as a part of it to understand the designing/development of neural net models using pytorch
- YOLO Object Detection
- Object detection is one of the basics of computer vision. It can be applied from smart lens' to autonomous driving vehicles. The YOLO object detection notebook here describes the design & implementation for multi-object detection from a perspective of an autonomous driving vechile.
- Pytorch - Most of the implementation specifically model training is done using the popular Pytorch framework
- COCO Dataset - COCO dataset was used in the training of image captioning project. It has thousands of images taken in real world scenarios with each image accompanied by set of 5 captions
- Numpy
- OpenCV