Hi,
I would like to show one of my research idea on Detecting objects and related information in Road Traffic. The system was implemented using Pretrained CenterNet HourGlass104 Keypoints 512x512 (Tensorflow Hub Model). The idea behind this system is to make a autonomous system for Cars that will help the drivers to understand the road conditions in a more better way and more detailed information in order to avoid accidents and even we can integrate these kind of systems in future autonomous driving system. The other possible arears where we can implement this kind of research is given below:
- Intelligent Driving Assistant for Drivers
- Help the new drivers to learn traffic situations in a more better way
- Government agencies can use to enforce rules and fines by detecting such strange activities.
This codes are implemented using Tensorflow Hub Pretrained Models.
I have tried to test this model on Indian Traffic Conditions which are given below, the demo is given below:
The source codes are available here: GitHub - rohitnarain24/Traffic-objects-detection-using-Tensorflow-Hub-Models
I hope this will help others to learn, how to use these pretrained models in real application areas and images.
Thanks