MahmudulAlam / Automatic-Identification-and-Counting-of-Blood-Cells

Machine learning approach of automatic identification and counting of blood cells (RBC, WBC, and Platelet) with KNN and IOU based verification.
https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098
GNU General Public License v3.0
133 stars 53 forks source link

Weights issue #4

Closed testchange closed 5 years ago

testchange commented 5 years ago

Hi,

Thanks for the awesome work!

I am trying to replicate your work but run into some issues with your weights.

I have downloaded your weights but it comes in different formats. For darkflow, the weights file is .weights but I cannot find this file type from your weights.

(1)Do you have .weights file?

I tried renaming your file tiny-yolo-voc-3c-3750.meta and tiny-yolo-voc-3c-3750.data-00000-of-00001 at file size, 63mb and 189mb, to .weights file but the weight just does not match with your architecture, with the error message below when I run this command, using the original darkflow github repo

flow --model cfg/tiny-yolo-voc-3c.cfg --load bin/tiny-yolo-voc-3c.weights

Note: I copy your tiny-yolo-voc-3c.cfg and rename your file to tiny-yolo-voc-3c.weights to run the above command.

Error message:

AssertionError: expect 63082056 bytes, found 63676889
MahmudulAlam commented 5 years ago

Hello. First of all, .weights is the weight file format of Darknet, not darkflow. So, you should not change the weight file extensions. Secondly, .cfg or configuration file contains the network architecture from which network is built so you should not change that extension too. Finally, the command line arguments are not implemented for blood cell detection. So, you may use an IDE to run the code. I will try to create a wiki providing step by step guideline of how to run the code beginning from cloning the repo as soon as possible.

MahmudulAlam commented 5 years ago

I have created this step by step guideline of how to run the code. Hope this helps.