Activities per year
Abstract
Modern optical flow methods are often composed of a cascade of many independent steps or formulated as a black box neural network that is hard to interpret and analyze. In this work we seek for a plain, interpretable, but learnable solution. We propose a novel inpainting based algorithm that approaches the problem in three steps: feature selection and matching, selection of supporting points and energy based inpainting. To facilitate the inference we propose an optimization layer that allows to backpropagate through 10K iterations of a first-order method without any numerical or memory problems. Compared to recent state-of-the-art networks, our modular CNN is very lightweight and competitive with other, more involved, inpainting based methods.
Original language | English |
---|---|
Publication status | Published - 4 Dec 2018 |
Event | 14th Asian Conference on Computer Vision: ACCV 2018 - Perth Western Australia, Perth, Australia Duration: 4 Dec 2018 → 6 Dec 2018 http://accv2018.net |
Conference
Conference | 14th Asian Conference on Computer Vision |
---|---|
Abbreviated title | ACCV 2018 |
Country/Territory | Australia |
City | Perth |
Period | 4/12/18 → 6/12/18 |
Internet address |
Keywords
- Optical Flow
- Energy Optimization
- Deep Learning
Activities
- 1 Poster presentation
-
Learning Energy Based Inpainting for Optical Flow
Patrick Knöbelreiter (Speaker)
5 Dec 2018Activity: Talk or presentation › Poster presentation › Science to science