Tensorflow and keras implementation of YOLO algorithm using the on-board camera of TX2.
Demo Link:https://youtu.be/rHyBtfIDGOc
The neural network was trained on Nvidia Titan X GPU.This model was later used with nvidia Jetson TX2 Board. Opencv was used to capture images .
1.Python 3.5
2.Tensorflow 1.5
3.Keras
4.Scikit Learn
5.Open CV 3.4.1
Broad Overview
|
Reduce Dimesions of image
|
probablity of box | box co-rdinates | Probability of each class
|
hard threshold
|
Non max-suppression
|
output
1.Open CV installation for ubuntu is pretty standard and done in same way as shown on their website.
2.Installation of Open CV on Jetson TX2 is a bit diffrent.Jetson Hacks has a pretty cool script on his Git repo.This enable open CV on GPU on the Jetson.
This was done using pre-trained model by darknet.Follow the instruction as follows.
Then run the test_on_cam.py file.
Uncomment if needed to run on Video.Add video to data folder in .avi format.
Download the dataset: https://www.dropbox.com/s/nli1ne8hzkzsyt6/NFPAdataset.zip?dl=0
Yolo-tiny is a lightweight option with pretty descent accuracy and does not require huge computation resources.Athough a correct and accurate implementation is th YOLO designed above.
- Siddharth Bhonge - Parser /Model - https://github.com/siddharthbhonge
-
Andrew Ng | Deeplearning.ai
-
Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi - You Only Look Once: Unified, Real-Time Object Detection (2015)
-
Joseph Redmon, Ali Farhadi - YOLO9000: Better, Faster, Stronger (2016)
-
Allan Zelener - YAD2K: Yet Another Darknet 2 Keras
-
The official YOLO website (https://pjreddie.com/darknet/yolo/)