/object-detection-hls

TensorFlow Object Detection using local HLS server

Primary LanguagePython

Traffic Detection HLS

Simple implementation of OpenSSD Mobilenet V1 in TensorFlow to manage HLS video stream.

Usage

To run it, ensure that you have all of the files necessary:

  • Video stream data
  • Model & class label file
  • Installed required packages

And then you can run

HLS Server:

node generator/index.js

Object Detector:

python main.py

Generating Video Files

As of now, you have to modify the inputFile in modify.js to the file that you specify. Currently, it's hardcoded to input.mp4. Then you can run

node modify.js

Code Structure

traffic-detection-hls
 ├── 📜main.py              # Main Script
 ├── 📂internal             # Internal Code
 │ ├── 📜detector.py
 │ ├── 📜logger.py
 │ ├── 📜model.py
 │ ├── 📜params.py
 │ ├── 📜utils.py
 │ └── 📜visualizer.py
 ├── 📂generator            # HLS Server
 │ ├── 📂streams
 │ ├── 📜modify.js          # Code to generate video stream data
 │ └── 📜index.js           # Code to run HLS Server
 ├── 📂models               # Tensorflow Models Stored
 │ └── 📜labels.txt         # Class labels
 └── 📜requirements.txt