/cv_product_recognition_project

Computer Vision project for object detection. Grocery items are detected on a store shelf from single model images using local invariant features and the Generalized Hough transform.

Primary LanguageJupyter Notebook

Requirements

Python with Anaconda distribution and jupyter.

The following packages are needed:

  • openCV 4.5.5 or higher
  • matplotlib
  • scipy
  • numpy
  • shapely

Description

The overall tasks are developed in product_recognition.ipynb.

The two class FeatureMatcher and MultipleInstanceMatcher, contained in matcher.py, are explained step by step in workflow.ipynb.

Report

You can read and download our report here.

You can read the project assignment here.

Result

Step A - Single Product Detection

Step B - Multiple Product Detection

Step C - Multiple Product Detection