CoreHist - A Correlation Filter based Nuclei Detector for Histopathological Images

 

CoreHist is a cloud-based web-server designed to implement correlation filter based detector. It is based on a novel method using kernelized correlation filter for nuclei detection in histopathology images. The proposed filter offers state of the art prediction capability of nuclei in histology images. You can upload histopathology images of your interest to detect nuclei in them. The detected nuclei present themselves as correlation peaks in the response image. A sample input and response image are shown below. You may need to threshold and post-process the output. 

Input Image


CoreHist










Response Image

If you use our work or want to refer to it, you can cite Asif Ahmed, Amina Asif, M. Arif, Nasir Rajpoot and Fayyaz ul Amir Afsar Minhas- "Correlation Filter based Detection of Nuclei in Histopathology Images" Journal of Medical Systems (2018) 42:7. 

The code for the webserver is available at: this github page

 

Apply CoreHist to your image here

(The input image should be of size 500x500 and 24 bits in color)

 

Select file:       
            
                                       


Developed and maintained by Biomedical Informatics Research Laboratory, Department of Computer and Information Sciences (DCIS), Pakistan Institute of Engineering and Applied Sciences (PIEAS), Islamabad, Pakistan.