GrumpyZhou/semseg

TUM Practical course: Hands-on Deep Learning for Computer Vision

Study on pixel-level semantic segmentation and instance-level semantic segmentation

Practical course: Hands-on Deep Learning for Computer Vision from computer vision chair of TUM.

Pixel-level semantic segmentation:

We reimplementated FCN-8s with minor modification (More details in our report)

Result:

• View CityScape online Benchmark result of our FCN8s.
• A sample of semantic segmentation
• 

Instance-level semantic segmentation

See report