Unet for instance segmentation

different problem need to solute by different neural network.

In my this project problem is need to computer "watching" and "understand" a video.

That mean, i need a neural network can find out each object from a image/video.

(Image from : https://www.researchgate.net/figure/Convolutional-neural-network-CNN-architecture-based-on-UNET-Ronneberger-et-al_fig2_323597886)