Glare Reduction - Image Processing

Introduction

Glare reduction is a step in Recovering glare images by Image processing method in Convolutional Neural Network (CNN).

You can use this at step: pre-processing data in CNN.

Glare reduction includes 3-mixed-filter by 4 steps:

1. Reduce-glare filter
2. Enhance contract: f = 1.6
3. Reduce-glare filter
4. Enhance contract: f = 1.4

More details at Documentation.pdf

Development

1. Put your glare image into examples folder.
2. In your terminal:

    python generate.py && python show_case.py

3. Check out your Generated_Image folder.

Filters' Description

1. Reduce-glare filter:

Include a 4-mixed-filter by 4 steps:

1. First polynomial function
2. Gamma correction: g = 0.75
3. Second polynomial function
4. Gamma correction: g = 0.8

2. First polynomial function:

For every pixel value within [0; 255], First polynomial function map it to another value so that:

• The intensity with value less than 100 will increase.
• The intensity with value greater than 100 will decrease.

Orange line
x is original intensity
y = f(x) is intensity after applying the first polynomial function

First polynomial function's expression:

3. Second polynomial function:

For every pixel value within [0; 255], Second polynomial function map it to another value so that:

• The intensity with value less than 160 will increase.
• The intensity with value greater than 160 will decrease.

Orange line
x is original intensity
y = f(x) is intensity after applying the second polynomial function

Second polynomial function's expression:

Citation

If this project is useful in your research, please cite it using the metadata in CITATION.cff or the citation format shown on the GitHub repository page.