Background Removal from Video using Deep Learning | TensorFlow | Semantic Segmentation | OpenCV

In this video, we are going to build a python application which would help in removing background from human videos using deep learning. For this purpose, we are going to a custom model trained on the human image segmentation task in the TensorFlow framework.

Timeline:
00:00 - Introduction
00:27 - Video samples for removing background
01:15 - Weight file of the model trained on the People Segmentation dataset
02:01 - Importing the required libraries
03:35 - Defining the global parameters
03:48 - Defining a function to create a directory
04:16 - Main function
04:26 - Seeding the environment
04:56 - Create a directory/folder to save the output videos
05:11 - Loading the trained model
06:22 - Video path
07:15 - Reading the video to get the frame height and width
08:12 - Defining the variable for saving the output video
09:10 - Reading the video frame
10:27 - Processing the video frame
12:33 - Predicting the binary mask
16:07 - Processing the binary mask
17:42 - Removing background from the original frame
20:30 - Apply a custom colour to the background
23:26 - Saving the custom background coloured frame in the output video
25:06 - Videos with background removed
26:42 - Ending

Support:

Follow Me:

[lucasantagata]

Incredible tutorial, thank you! Could you share the links for the video that you used, please?

[burakkaratasz]

bro how i import this app to in android app ?