This repo contains all the computer vision projects that i have done for completion of the academic courses as well as for fun. The following projects along with their description are in this repository :
-
Hand Signs Images Classification : Dataset- Pictures of Hand Gestures showing digits from 1 to 5. Objective - To recognise digits from the images of hand signs How : Used Convolutional Network Architecture and AdamOptimizer with minibatches. Technologies : Tensorflow, pandas, image library of python Academic project completed during deeplearning.ai course.
-
Happy House Project : Dataset - Pictures of faces of people. Goal was to classify faces in two categories : happy and unhappy. if network says happy then only the person will be allowed to enter the house. Used Convonet Architecture implemented with Keras framework.
-
Dog Breed Classification : This is a featured competition at Kaggle. I participated in that competition and contributed this project. Implemented pretrained vgg19 neural network architecture for prediction of 120 breed classes.
-
Resnet Implementation : Implemented Residual Network Architecture using Keras and Tensorflow
-
Autonomous driving application : Car Detection with Yolo - In this project, Implemented Yolo paper with exact details written in the paper. This project was completed as a part of deep learning Specialization.
-
Face Recognition for Happy House : Implemented Facial Recognition system used in FaceNet with Inception Network. Implemented triplet loss, Face Verification and Face Recognition with pretrained weights of Inception Network.
-
Neural Style Transfer : Implemented Neural Style Transfer with VGG19 Architecture. Implemented Content Matrix, Style Matrix, and Content cost and Style cost to ultimately find the Cost function for the Network, which is linear combination of the former two costs.