search

ignaciorlando / red-lesion-detection

This code implements a red lesion detection method based on a combination of hand-crafted features and CNN based descriptors. Our paper is under revision now, so please do not use this repository until we release the paper.
47 stars 14 forks source link

tsne_representation can not run #15

Open sufenger opened 5 years ago

sufenger commented 5 years ago

Incorrect use + The matrix dimensions must be consistent.

script_tsne_representation (line 79) Is(:, :, :, iterator : iterator + size(imdb_cnn.imdb.images.data{i},1) - 1) = uint8((windows.imdb.images.data{i} + detector.net.meta.trainOpts.dataMean));

and i see "detector.net.meta.trainOpts.dataMean" is 32323,and "windows.imdb.images.data{i}"is 32323*139"(There are 139 candidates here.) i not konw how to do wite it , if you can help me ,I will be grateful.thank you .

sufenger commented 5 years ago

and cnn_filename is in the "red-lesions-detection-model " folder “classbalancingsoftmax-lr=0.05-ceps=0.0001-lreps=0.01-wd=0.005-batch=100-N=10-dp=0.01-fc=128.mat。“

imdb_hand_crafted_filename 、imdb_cnn_filename 、windows_filenames ，These three files are in My test picture folder’s “MA_candidates_data”.

I don't know if this setting is correct, and whether my question is related to my file. I will be very grateful for your answer. Looking forward to your reply！thank you again.