This repository contains solutions of tasks made during the Pattern Recognition academic course.
Topics:
- Hu Moments - calculation of Hu Moments for three letters and transformations of them (shifting, rotation, font change, resize, adding noise)
- Hough transform - determination of Hough space for picture of few circles and its distortions
- CS and DCT - checking how removing different parts od DCT spectrum affects image reconstruction with IDCT, also checking different strategies to choose pixels for CS image reconstruction, code based on http://www.pyrunner.com/weblog/2016/05/26/compressed-sensing-python/ with usage of library pyLBFGS from https://bitbucket.org/rtaylor/pylbfgs/src/master/
- Convolutional neural networks - using transfer learning on MobileNet v2 from TensorFlow to make neural network classifying pictures into 3 (or 4) categories
- Mini Inception Net - using convolutional neural network to classify images from CIFAR-10
- GAN - simple GAN producing fake images