Tire detector using HOG

This code is part of my project for the distributed systems subject on the master's degree in computer and systems engineering

More on HOG detectors:

https://lear.inrialpes.fr/people/triggs/pubs/Dalal-cvpr05.pdf

The algorith is based on the one showed in this article: https://towardsdatascience.com/hog-histogram-of-oriented-gradients-67ecd887675f

Steps

Get dataset

1. Get images with tires

2. Label tires in images

3. Extract only tires

4. Get images of things other than tires

Code

1. Load image

2. Resize image in 64 x 64

3. Calculate gradients

4. Divide in 8x8 cells

5. Calculate histograms

6. Normalize 2 x 2 blocks

7. Train classifier (p.e. SVM)

8. Create a menu to:

a. Train

b. Test

c. Predict

9. ?

10. Profit

Notes

For training, I need to input a folder path, it should generate a json file with training results. For testing I should pass a folder path and a json file with trained weights, and to predict i should pass an image path and a path for the json trained weights.

Compiling instructions

To compile this program, I need to have cmake installed

mkdir build
cd build
cmake ..
make

It generates a build/TireDetector executable file, then I can use with

build/TireDetector /path/to/images