Automatic evaluation of tumour budding in immunohistochemically stained colorectal carcinomas and correlation to clinical outcome [Dataset]https://doi.org/10.11588/data/XJAOC4Weis, Cleo-AronheiDATA2018-08-202018-08-31T06:56:58Z<b> Data used for the implementation of the proposed tumor budding detection</b><br />
In the publication “Automatic evaluation of tumour budding in immunohistochemically stained colorectal carcinomas and correlation to clinical outcome” we described a multistep approach to detect tumor buds in immunohistochemically stained images: . <br />
Step 1: Color and size based segmentation. <br />
Step 2: Validation of the detected objects (proposals) by a spatial clustering and a convolutional neural network (MatConvNet by A. Vedaldi et al. [1]). <br />
<p><img src="https://heidata.uni-heidelberg.de/api/access/datafile/1772?imageThumb=400&pfdrid_c=true"></p>
<br />
The Matlab-Code for the project is available on <a href="https://github.com/catweis/Automatic-evaluation-of-tumour-budding-in-immunohistochemically-stained-colorectal-carcinomas-">GitHub</a>.
<br />
The data for the CNN-training and validation are presented as .mat-file. It contains a struct element with the images in a 4D-matrix, the label (“bud” and “no bud”) and a set (“training” and “validation”).<br />
Please refer to the "Terms" tab below for usage and reproduction terms.<br />
<b> References:</b><br />
1. Vedaldi, A., K. Lenc, and A. Gupta. MatConvNet: CNNs for MATLAB. 2015; Available from: http://www.vlfeat.org/matconvnet/.
Medicine, Health and Life Sciences2018-08-202018-06-26Licensed under a <a href='https://creativecommons.org/licenses/by-sa/4.0/'>Creative Commons Attribution-ShareAlike 4.0 International License.  <img src='https://i.creativecommons.org/l/by-sa/4.0/80x15.png' alt='CC by sa' /></a>