Tensorflow-Object-Detection-Inference

This is minimal script to run inference on the test images and dump JSON results with detected classes for each image

Directory Structure:

├── LICENSE
├── Object_Detection_Inference_Minimal_V1.py
├── README.md
├── object_detection
│   ├── Inference_Graph
│   ├── Label_Map
│   │   └── labelmap.pbtxt
│   ├── Test_Images
│   │   ├── cat.jpg
│   │   └── dog.jpg
│   └── utils
│       ├── __pycache__
│       │   ├── label_map_util.cpython-38.pyc
│       │   └── ops.cpython-38.pyc
│       ├── label_map_util.py
│       └── ops.py
└── output.json

Setup Steps:

• Copy exported inference graph files and folders into object_detection -> Inference_Graph directory.
• Copy your labelmap.pbtxt into object_detection -> Label_Map directory.
• Copy test images into object_detection -> Test_Images directory.
• Run the inference script

foo@bar:~$ python Object_Detection_Inference_Minimal_V1.py

• The output is dumped into the top directory with a name output.json