Computer Vision - Shaastra Trainer
Computer Vision
Computer vision is a field of artificial intelligence (AI) that enables computers and
systems to derive meaningful information from digital images, videos and other visual
inputs and take actions or make recommendations based on that information.
Abstract
This session will start from the basics of Neural Networks, Convolutional Neural
Networks and go diving deep into Semantic Segmentation, GANs etc. We will also cover
how to start a Computer Vision project from scratch and give insights into it. At the
end of the workshop you will be able to start your own Computer Vision Project and
compete in Competitions :)
Duration
2 sessions : 3 hours each
Session 1 Neural Networks, Activation Functions and Optimizers, What is an
Image?, Convolutional Neural Networks, Model Architecture on CNNs
Session 2 Application of Computer Vision, Python Libraries, First Computer Vision
Project, How to read a research paper
Pre-requisites
Basic knowledge on how Neural Networks work would be recommended. The session
will be going over that topic very briefly
Detailed Plan
The whole workshop will span for 2 days and consist of 2 sessions with 9 modules in
total
Session 1
Module 1 : Neural Networks
● Forward Propagation
● Backward Propagation
● Gradient Descent
● Normalisation
Module 2 : Activation Functions and Optimizers
● Activation Functions like ReLU, Sigmoid and when to use them
● Optimizers and concept of Momentum and Hyperparameters
Module 3 : What is an Image?
● Pixels and Channels
● Colour and Grayscale Image
● Data Augmentation
Module 4 : Convolutional Neural Networks
● Filters and How they work
● Padding, Strides and Pooling Layer
● Why are CNN’s so good for Images
Module 5 : Model Architecture on CNNs
● Basic Models : VGG, AlexNet, LeNet etc
● Models with Skip Connections : ResNet, DenseNet etc
● Inception Networks and other popular models
Session 2
Module 6 : Application of Computer Vision
● Image Classification
● Semantic Segmentation
● Object Detection
● Generative Adversarial Networks (GAN)
Module 7 : Python Libraries
● What are Python Libraries and why is it useful?
● Numpy
● Matplotlib
● Other useful libraries like OpenCV, Pandas etc
Module 8 : Your First Computer Vision Project
● How to start the project
● Steps to follow while doing the project
● How to find errors and debug them
Module 9 : How to read a research paper