Detecting Multiple objects in a video using Single Shot Multibox Detector

Weight Files are splitted as

• ssd300_mAP_77.43_v2.pth.000
• ssd300_mAP_77.43_v2.pth.001
• ssd300_mAP_77.43_v2.pth.002
• ssd300_mAP_77.43_v2.pth.003
• ssd300_mAP_77.43_v2.pth.004

Join weight files Here

Read more about SSD here

Click on this image to see demo from SSD:

Check out the
virtual_platform_windows.yml file

Create a virtual platform.
conda env create -f virtual_platform_windows.yml

[]

updateobject_detection.py
reader = imageio.get_reader('video.mp4')

Training is simple but GPU CUDA is mandatory otherwise it will take months to train on a CPU:

#Get the dataset:
(http://host.robots.ox.ac.uk/pascal/VOC/index.html)

Created by:
Anubhav Shukla