mjdelta/yolo2-keras

detecting Red Blood Cell with yolov2

yolo2-keras

How to use this repo.

• Train on your RBC datasets.
• Load my model to detect Red Blood Cell.
• Train step by step.

Train on your RBC datasets

First, please download the weights I have trained here(https://drive.google.com/open?id=1-M84cxZIyVeBUfxat0MF7Wn7hmLHUmsk) and put both in a directory named "models/". Its full path should be "models/tiny_yolo_backend.h5" and "models/tiny_yolo.h5".

Second, please run train.ipynb

Load my model to detect Red Blood Cell

Please run load_model.ipynb

Train step by step.

Step 1：Download the tiny_yolo_backend.py here(https://drive.google.com/open?id=1-M84cxZIyVeBUfxat0MF7Wn7hmLHUmsk) and put it in a directory named "models/". Its full path should be "models/tiny_yolo_backend.h5".

Step 2: In shell, run python train.py -c config.json.

Step 3: In shell, run python test.py -c config.json -w models/tiny_yolo.h5.

Here's the detected sample.

Please enjoy it! :)

yolo2-keras's Issues

where to download weight files

i can not download the weight files.
is it valid yet?

or i am blocked by my country's firewall?
would u please send the weight files to my E-mail: [email protected]

Full Yolo Detect

Have you tried the full yolo for detecting or used the tiny yolo for other dataset ? I still find the confidences of the boxe are 0.5.

SOS weight files

i can not download the weight files.
is it valid yet?

or i am blocked by my country's firewall?
would u please send the weight files to my E-mail: [[email protected]]