The SSD Inception V2 network can be used to detect a number of objects specified by a particular training set. This model in particular can detect the following produce/items usually found in grocery stores:
- bell peppers
- olive oil
The provided Makefile does the following:
- Downloads a trained model
- Downloads test images
- Compiles the network using the OpenVINO Model Optimizer
- There is a python example (run.py) which runs an inference for all of the test images to show how to use the network with the OpenVINO toolkit.
This network is based on the TensorFlow Object Detection API SSD Inception V2 model. The model was modified, trained, and saved in order to be compatible with the OpenVINO toolkit.
The example runs inference using a video stream taken from an existing camera device. Bounding boxes and their associated probabilities and classifications are rendered on frames of the video stream and presented on a OpenCV frame.
Provided Makefile describes various targets that help with the above mentioned tasks.
Runs a sample application with the FP16 network. Users must plug in their Intel Neural Compute Stick 2 in order to successfully run this application.
ssd_inception_v2_food.py script which takes in a video stream from an existing camera, runs inference on each frame, and renders the output to the user.
ssd_inception_v2_food.py script with the FP32 network. Note that this application will only run successfully on CPUs that OpenVINO supports.
ssd_inception_v2_food.py script with the FP16 network. Users must plug in their Intel Neural Compute Stick 2 in order to successfully run this application.
Trains a SSD Inception V2 model using the Tensorflow Object Detection API given an
JPEGImages folder containing .xml and .jpg images, respectively, for training. Training is not necessary since the sample will download a pre-trained model. This option allows for the user to further refine the SSD Inception V2 model if they so desire.
Compiles model from trained SSD Inception V2 model. This is not the same as
make compile_model since we are compiling the model from
make train on your local computer.
Shows makefile possible targets and brief descriptions.
Makes the follow items: deps, data, compile_model.
Compiles the trained model to generate a OpenVINO IR file. This file can be loaded on the Neural Compute Stick for inferencing.
Downloads the trained model.
Downloads and prepares a trained network for compilation with the OpenVINO toolkit.
Checks required packages that aren’t installed as part of the OpenVINO installation.
Uninstalls requirements that were installed by the sample program.
Removes all the temporary and target files that are created by the Makefile.