Computer Vision Roadmap

This repository contains the process I am following to learn Computer Vision. I am starting from the very beginning, so the level of projects would behind from most basic and hopefully become somewhat advanced as this repository evolves. Follow along if you are keen to explore the fantastic projects you can make using Computer Vision capabilities along with Deep Learning.

Requirements for Face Detection

1. Download Anaconda
2. Download the Installation files from the Face Detection folder in the repository
3. Mac and Linux users, fire up the terminal. Windows users can open Anaconda Prompt after downloading Anaconda
4. Navigate to the folder with installation files using cd command
5. Run this command for setting up dependencies on Mac: conda env create -f virtual_platform_mac.yml
6. Windows and Linux users can run the same command as conda env create -f virtual_platform_windows.yml and conda env create -f virtual_platform_linux.yml , respectively.
7. This will setup a virtual_platform on your system with all the dependencies
8. Go to Anaconda application on your machine and on the top left where you can see root, select Virtual_Platform from the drop down list
9. Download the 2 xml files, namely haarcascade_eye.xml and harcascade_frontalface_default.xml to a known location on your machine.
10. Change the reference location of these 2 xml files while running the face_detection.py script

Requirements for Object Detection

1. Download Anaconda
2. Download the Installation files from the Face Detection folder in the repository
3. Mac and Linux users, fire up the terminal. Windows users can open Anaconda Prompt after downloading Anaconda
4. Navigate to the folder with installation files using cd command
5. Run this command for setting up dependencies on Mac: conda env create -f virtual_platform_mac.yml
6. Windows and Linux users can run the same command as conda env create -f virtual_platform_windows.yml and conda env create -f virtual_platform_linux.yml , respectively.
7. This will setup a virtual_platform on your system with all the dependencies
8. Go to Anaconda application on your machine and on the top left where you can see root, select Virtual_Platform from the drop down list
9. Take all the files inside the Object Detection folder from Github Repository and upload them onto jupyter notebook home. ( all the files except Output, obviously!)
10. Now run the cells one-by-one in the Object Detection jupyter notebook.
11. You will see an output file being generated. Open the file and enjoy the video!

Requirements for Real time Object Detection

1. Download Anaconda
2. Download the Installation files from the Face Detection folder in the repository
3. Mac and Linux users, fire up the terminal. Windows users can open Anaconda Prompt after downloading Anaconda
4. Navigate to the folder with installation files using cd command
5. Run this command for setting up dependencies on Mac: conda env create -f virtual_platform_mac.yml
6. Windows and Linux users can run the same command as conda env create -f virtual_platform_windows.yml and conda env create -f virtual_platform_linux.yml , respectively.
7. This will setup a virtual_platform on your system with all the dependencies
8. Go to Anaconda application on your machine and on the top left where you can see root, select Virtual_Platform from the drop down list
9. Download data by following the steps given here: cifar-dataset
10. Save the data in a folder in Jupyter notebooks. Name the folder 'data_for_gan'

Results

1. Face Detection

2. Real- Time Object Detection Output Video

3. Generative Adversarial Network

Books

• Deep Learning with Python
• Deep Learning by Ian Goodfellow and Yoshua Benjio and Aaron Cornville

Courses

• Deep Learning and Computer Vision A-Z

Interesting Blog Posts

• Everything you need to know about Computer Vision - Link
• Viola Jones Algorithm for face detection - Link
• OpenCV3 Object Detection - Link
• SSD - Link
• GANs - Link

Research Papers

• SSD
• Viola Jones
• GANs