A python application developed to demonstrate object recognition using pre-trained models from TensorFlow Model Zoo.
-
Clone the repo and create a virtual environment using something like virtualenv.
-
Install all dependencies using
pip install -r requirements.txt -
Download these models and extract them
SSD Inception V2
SSD MobileNet V2
Faster RCNN Inception V2 -
The directory structure should look like
object_recognition_poc/
├── coco_labels.txt
├── requirements.txt
├── images
│ ├── car and pedestrian.jpg
│ ├── desk (1).jpeg
│ ├── desk.jpg
│ ├── furniture.jpg
│ ├── puppy.jpg
│ └── vehicles.jpg
├── models
│ ├── faster_rcnn_inception_v2_coco_2018_01_28
│ │ └── frozen_inference_graph.pb
│ ├── ssd_inception_v2_coco_2018_01_28
│ │ └── frozen_inference_graph.pb
│ └── ssd_mobilenet_v2_coco_2018_03_29
│ └── frozen_inference_graph.pb
└── src
├── annotations.py
├── gui.py
├── main.py
└── object_detector.py - Run
src/main.py
Tested only on Ubuntu 18.04, might not work on Windows.