Computer Vision Tasks

This repository contains three tasks that have been written in Python3 as part of the coursework for the Computer Vision course at the University of Bath.

The 3 tasks are:

1. Angle Finding
2. Template Matching
3. Scale-Invariant Feature Transform (SIFT)

Table of Contents

1. Organizing Image Data
   1. Angle Finding
   2. Template Matching & SIFT
2. Run Instructions
3. Settings
4. License

Image Data

If there is no image data provided in the /data folder or you want to try out your own images with the algorithms, please do the following:

Task 1 - Angle Finding

Please place your images into src/data/task_1/test_images, and update the list.txt file accordingly.
The format for the list.txt file is <image_name>.<file_type (e.g., png)>,<angle in degrees> (without the <>)

Task 2 | Task 3 - Template Matching & SIFT

Please place your test images into src/data/task_2_3/test_data/images, the annotations for the test data into src/data/task_2_3/test_data/annotations and the training data templates into src/data/task_2_3/training_data.

2. How to Run

There are 2 possible ways how to run the code:
(assuming you are one folder above the /src directory)
1. Call the tester.py via python3 tester.py. You will be asked which algorithm to run. The tester will automatically fetch all data from the data folder and run the algorithm(s).
2. Call the run_algorithm.py if you want to run the algorithms with single images. For that, open your console and run the file via python3 run_algorithm.py <path_to_test_image>.

3. Settings.ini

Feel free to toggle the parameters in the settings.ini located in the /src folder to produce different output/logging/etc. You can also play around with the pyramid parameters that were used for the template matching.

4. License

All code is available under the MIT License. See LICENSE for the full license text.