Deep Learning - Segmentation and Object Detection Part 2
In this video, we discuss ideas on how to improve on image segmentation. In particular skip connections as used in the U-Net have been applied very successfully here. Also, we, look into other advanced methods such as stacked hourglasses, convolutional pose machines, and conditional random fields in combination with deep learning.
Additional References
nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation
X-ray-transform Invariant Anatomical Landmark Detection for Pelvic Trauma Surgery
Further Reading:
A gentle Introduction to Deep Learning