Scalable Full Flow with Learned Binary Descriptors
| Autor: | Thomas Pock, Patrick Knöbelreiter, Gottfried Munda, Alexander Shekhovtsov |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Conditional random field
FOS: Computer and information sciences Matching (graph theory) Computer science Computation Computer Vision and Pattern Recognition (cs.CV) Optical flow ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Computer Science - Computer Vision and Pattern Recognition Binary number 020207 software engineering 02 engineering and technology Convolutional neural network Scalability 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing State (computer science) Algorithm |
| Zdroj: | German Conference on Pattern Recognition Lecture Notes in Computer Science Lecture Notes in Computer Science-Pattern Recognition Lecture Notes in Computer Science ISBN: 9783319667089 GCPR |
| ISSN: | 0302-9743 1611-3349 |
| DOI: | 10.1007/978-3-319-66709-6_26 |
| Popis: | We propose a method for large displacement optical flow in which local matching costs are learned by a convolutional neural network (CNN) and a smoothness prior is imposed by a conditional random field (CRF). We tackle the computation- and memory-intensive operations on the 4D cost volume by a min-projection which reduces memory complexity from quadratic to linear and binary descriptors for efficient matching. This enables evaluation of the cost on the fly and allows to perform learning and CRF inference on high resolution images without ever storing the 4D cost volume. To address the problem of learning binary descriptors we propose a new hybrid learning scheme. In contrast to current state of the art approaches for learning binary CNNs we can compute the exact non-zero gradient within our model. We compare several methods for training binary descriptors and show results on public available benchmarks. GCPR 2017 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|