Object detection example using CNTK and Python
This example is modified version of predefined CNTK Example of Image detection which can be found at http://github.com/Microsoft/CNTK.
By running Nokia3310_detection.py the code will do the following:
- Train and create CNTK based model whih can detect NOkia3310 on an image.
- By providing test images the model can be evaluated and tested.
- The example uses minimal python code, needed to run object detection by using FasterRCNN.
- The model is not trained from zero, it is based on AlexNet pre trained model.
To run Nokia3310 object detection example you need a CNTK 2.5, Python 3.5 environment. In order to install CNTK2.5 you have to install Intel Math Kernel Library (MKLML) from this location https://github.com/intel/mkl-dnn/releases. Anyhow in order for proper installation of CNKT consult to the official site.
Beside the basics requiremens you need to install the following additional packages:
pip install opencv-python easydict pyyaml future pillow
Also, code uses prebuild Cython modules for parts of the region proposal network (see utils/cython_modules). This will only work on Windows with Python 3.5 at the moment. You'll need to compile the cython module binaries for your specific platform (OS & Python version) to have this work on another environment.
If you want to use the debug output you need to run pip install pydot_ng (website) and install graphviz to be able to plot the CNTK graphs (the GraphViz executable has to be in the system’s PATH).
The example uses the pre-trained AlexNet model which can be downloaded by running the following Python command from the PretrainedModels folder:
python download_model.py
To train and evaluate a detector run
python Nokia3310_detection.py
In order to change DataSet you have to provide the images and data. More information about how to prepare image and data can be found at http://bhrnjica.net
Changing base model is not supported.