This was a learning project pursued to step into the world of Machine Learning and Computer vision to be specific.
This project was divided into many parts which include:
- CatVsDogs model - A model that classifies an image as a cat or a dog. It cleared the general concepts and terminologies of computer vision. Different architectures (different settings of convolutional layers) were trained and their corresponding performance was recorded to find the optimum settings. README
- Trying YOLOv3 to get familiar with the application of the fastest object detection methods available. Learning how dows YOLO algorithms work and how are they faster than others.
- Making a web based application on Computer Vision using flask framework. This was the major part of the project. YOLOv5 is implemented to detect, classify and put bounded boxes in a live stream of images (coming from webcam). The output is rendered on a simple web page, which also includes live updating list of objects. Another feature added was that the background of the list changes colour to RED when a cell-phone is detected in the stream.README
(Separate descriptions and screenshots are added to respective folders)