HFUT-CV-LAB-2022

Implementation for labs from HFUT Computer Vision course

There are three labs during this course, as shown below:

• Lab1 : Line detection based on Hough Transform
• Lab2 : Image segmentation based on any methods in CV
• Lab3 : Image classification based on any methods in CV

Description:

Platform

• Windows 10
• Pycharm 2022.1
• Python 3.8

lab1

Implement line detection algorithm based on hough transform.

These libraries are needed:

• opencv-python 4.5.5.64
• numpy 1.22.3

python main.py

Test images in assets folder Results in results folder

lab2

Implement image segmentation algorithm based on meanshift

These libraries are needed:

• opencv-python 4.5.5.64
• numpy 1.22.3
• scipy 1.4.1

python meanshift.py

Test images in assets folder Results in results folder

lab3

Implement image identification algorithm based on CNN

• CNN(LeNet-5) for MNIST datasets
• modified LeNet-5 for CIFAR-10 datasets

These libraries are needed:

• matplotlib 3.5.1
• numpy 1.22.3
• sklearn 0.0
• tensorflow 2.10.0
• keras 2.10.0

Train model based on MNIST datasets and output accuracy

python cnn.py mnist --option train

Load pre-trained model, test and output accuracy

python cnn.py mnist --option test

Train model based on CIFAR-10 datasets and output accuracy

python cnn.py cifar-10 --option train

Load pre-trained model, test and output accuracy

python cnn.py cifar-10 --option test