vishal9629 / capture-it-openly Goto Github PK

# Capture-it-openly 📷 Team Bytezz ✨ presents the project 💻 Capture it Openly made during 48 hours of Hackathon- Electrothon3.0 conducted by SPEC, NIT Hamirpur under the Human Removal Foreground Challenge(Organisational Problem Statement from Spyne.ai)

What we Did?

We have developed an effective solution to facilitate picture 📷 capturing so that people can freely capture and edit according to their own wishes by removing the unwanted part without reducing the quality image. This has been achieved using Semantic Segmentation,Floodfill Function and GANS Model

Our Approach:

We trained a semantic segmentation model that automatically creates the mask over the subjects of the image that improves user-experience

Applied floodfill function on our model’s output that helps the user to keep the main subject of the image.

Currently,we are working on data set of CGANS Model

**References 📁 we have taken:

Semantic Segmentation Study Source

Convolutional Networks for Biomedical Image Segmentation

Semantic Segmentation