Real time Video Segmentation
Event details
Date | 29.08.2019 |
Hour | 14:00 › 16:00 |
Speaker | Evann Courdier |
Location | |
Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Martin Jaggi
Thesis advisor: Dr. François Fleuret
Co-examiner: Dr. Mathieu Salzmann
Abstract
Image semantic segmentation, the process of assigning a class label to every pixel in an image, is a critical component for numerous applications from Medical Imaging to Video Surveillance to Autonomous Driving. Some of these applications require real-time outputs of the segmentation systems to be usable. However, these systems are usually relatively slow and require high computing power.
This projects focuses on bringing image segmentation to low computing power devices that need to run image segmentation real-time, as for autonomous driving and drones. In particular, few works leverage the temporal coherence of successive images inside a video to produce a more efficient network.
Background papers
Encoder-Decoder with Atrous Separable Convolution for Semantic,Image Segmentation (DeepLabv3+), by Chen, L.-C., et al.
BiSeNet:Bilateral Segmentation Network for Real-time Semantic Segmentation, by Yu, C., et al.
Low-latency video semantic segmentation, by Li, Y. et al.
Exam president: Prof. Martin Jaggi
Thesis advisor: Dr. François Fleuret
Co-examiner: Dr. Mathieu Salzmann
Abstract
Image semantic segmentation, the process of assigning a class label to every pixel in an image, is a critical component for numerous applications from Medical Imaging to Video Surveillance to Autonomous Driving. Some of these applications require real-time outputs of the segmentation systems to be usable. However, these systems are usually relatively slow and require high computing power.
This projects focuses on bringing image segmentation to low computing power devices that need to run image segmentation real-time, as for autonomous driving and drones. In particular, few works leverage the temporal coherence of successive images inside a video to produce a more efficient network.
Background papers
Encoder-Decoder with Atrous Separable Convolution for Semantic,Image Segmentation (DeepLabv3+), by Chen, L.-C., et al.
BiSeNet:Bilateral Segmentation Network for Real-time Semantic Segmentation, by Yu, C., et al.
Low-latency video semantic segmentation, by Li, Y. et al.
Practical information
- General public
- Free