Laboratory for Knowledge Inference in
Medical Image Analysis

  Home About us Research The Team News Code and Data  

Pathology Images: KIMIA Path960

The KIMIA Path960 is a dataset that was proposed in the following paper:

A Comparative Study of CNN, BoVW and LBP for Classification of Histopathological Images

Meghana Dinesh Kumar, Morteza Babaie, Shujin Zhu, Shivam Kalra, and H.R.Tizhoosh; The 2017 IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2017), Honolulu, Hawaii, USA from Nov. 27 to Dec 1, 2017. 

This paper introduced a new dataset of histopathology images "KIMIA Path960". From a collection of more than 400 whole slide images (WSIs) of muscle, epithelial and connective tissue etc., we selected 20 scans that "visually" represented different texture/pattern types (purely based on visual clues). We manually selected 48 regions of interest of same size from each WSI and downsampled them to 308x168 patches. Hence, we obtained a dataset of 960(=20x48) images. The images are saved as color TIF files although we do not use the color information (i.e., the effect of staining) in our experiments.


:: Download the paper

:: Download the KIMIA Path960 image dataset [129 MB]

:: Our Sponsors

KIMIA Path 960 Image Dataset

Sample images for 20 classes in the KIMIA Path960 image dataset; the patches have been manually selected from a large set of histo-pathology scans.



© KIMIA Lab, University of Waterloo, Canada, 2013-2019 - Using Artificial Intelligence to Infer Knowledge from Medical Images