single-class-object-detection

The codes include data and data.rec. Based on YoLo algorithm and 4 existed model to detect single-class objects.In the code, it builds CNN body and Loss function in MXNET. Also, this code has the function to check tensorboard.

data

You can find the Data.rec file model and para file in those links:

https://drive.google.com/file/d/17upbN20WpzLEGCzFVWJJ1gHlAbXrsopB/view?usp=sharing

https://drive.google.com/open?id=17upbN20WpzLEGCzFVWJJ1gHlAbXrsopB

how to train and test

To train the code, you should use:

python run_train.py

And GPU is needed with Mxnet,maybe you need to change the path of data you load

final test could be used to test the model as below:

The revised one which will have the function to detect every object in one picture will be uploaded recently.

end