Daily CVPR - Wednesday

Morning: 09:00AM to 10:30 AM Semantic Segmentation Oral session O3-1B-13: Instance-Aware Semantic Segmentation via Multi-Task Network Cascades Page 33 of the Pocket Guide and presentation 13 on CVPR’s website Spotlight session S3-1B-20: The Cityscapes Dataset for Semantic Urban Scene Understanding Page 33 of the Pocket Guide and presentation 20 on CVPR’s website Afternoon: 1:45PM to 3:20PM Video Understanding Oral session O3-2A-1 Deep Hand: How to Train a CNN on 1 Million Hand Images When Your Data Is Continuous and Weakly Labelled Page 36 of the Pocket Guide and presentation 1 on CVPR’s website Video Analysis 2 Spotlight session S3-2A-11 Optical Flow With Semantic Segmentation and Localized Layers Page 36 of the Pocket Guide and presentation 11 on CVPR’s website 2 CVPR Daily: Wednesday For Today, Wednesday 29 David’s Picks David Pfeiffer is Principal Engineer at Daimler AG . He holds a Ph.D. in Computer Science from Humboldt University of Berlin and works in R&D for real-time computer vision and machine learning in the field of autonomous driving. David accepted to share his picks for today with CVPR Daily . Here are David’s picks for Wednesday 29. Don’t miss them! “ Multi-task network is really worth looking at: solving multiple problems looking at one only image ” “ A great dataset to work with, exceeding everything there is in the field of pixel-wise, accurate semantic labels” “ 1 million hand images were not well labelled and the question is how to make sense of it. We do not work on hand gestures, but images are continuous, like video, and there might be good ideas that I might use in the automotive industry” “ This is something we are working on right now…”