Hybrid Images

This project is an implementation for Hybrid Images based on Gaussian low and high pass filter, coded in Python language.

We implement hybridizing of two images in two different ways, one is conventional convolution, the other is FFT accelerated convolution.

Here we represent 6 pair images to show the hybrid result.

Project Structure

• Runnable Python source files are HybridImage_Gray.py and HybridImage_RGB, which are implemented on Gray Scale and RGB Scale. JUST CLONE THE REPOSITORY AND RUN IT!
• Images to be hybridized are in dataset_hybrid_images directory, which contains 6 image pairs.
• hybrid_result_gray and hybrid_result_rgb directories are outputs of the two runnable Python scripts, which represent our result of hybrid images.

Algorithm Process of Making Hybrid Image

• Extract Low Pass Frequency
• Extract High Pass Frequency
• Combine Low Pass and High Pass Frequency

Dependency

• OpenCV 3.4.1

References

• [1] Oliva, Aude & Torralba, Antonio & Schyns, Philippe. (2006). Hybrid images. ACM Trans. Graph.. 25. 527-532. 10.1145/1141911.1141919.

Author Info

LeoHao (XMU-CS)

Date

2020.11.03