HOG Feature Extraction for Human Detection using KNN

An academic project for Computer Vision (CS-GY-6643) at NYU Tandon with Prof Edward Wong.

Objective: To classify a test image as Human or Non-Human by using HOG feature extraction and KNN clustering for classification against a small training set of 20 labeled images.

For more details: Prompt & Report

Execution Instructions

1. Ensure python3 is setup on your system.
2. Download the python and image files into a directory. Please keep the image folder structure the same or the code will not work.
3. Enter above directory and pass a test image as a parameter to run the program as follows:

python3 hog.py './Test images (Neg)/00000003a_cut.bmp'

Contributors

Suyash Soniminde - sys8910@nyu.edu

Viha Gupta - vg2237@nyu.edu