The theory of this project based on the work of Professor Pappas's former student S. Derlin Babacan in 2005. The student took the research result away and applied for a patent and set up his own company but Professor Pappas always wanted to know how the algorithm was implemented, so he gave me the task to think about.

 

At present, the research work has been basically successful. This image segmentation algorithm is based on Gibbs Random Field theory and Python code. It can segment video image reasonably and recognize foreground and background. It may play an important role in the field of computer vision. 

​

The Following images show single image segmentation:

Original image vs My algorithm vs K-mean

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

​

​

The following GIFs and videos show the video segmentation results.

​