Joint unsupervised disparity and optical flow estimation of stereo videos with spatiotemporal loop consistency = 시-공간 루프 일관성을 이용한 비지도학습 기반 스테레오 비디오 깊이 및 광학 흐름 추정

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 56
  • Download : 0
Most existing deep learning-based depth and optical flow estimation methods require the supervision of a lot of ground truth data, and hardly generalize to video frames, resulting in temporal inconsistency (flickering). In this paper, I propose a joint framework that estimates disparity and optical flow of stereo videos and generalizes across various video frames by considering the spatiotemporal relation between the estimated disparity and flow without supervision. To improve both accuracy and consistency, I propose a loop consistency loss which enforces the spatiotemporal consistency of the estimated disparity and optical flow. Furthermore, I introduce a video-based training scheme using the convolutional Long Short-Term Memory (c-LSTM) to reinforce the temporal consistency. Extensive experiments show proposed methods not only estimate disparity and optical flow accurately but also further improve spatiotemporal consistency. This framework outperforms the current state-of-the-art unsupervised depth and optical flow estimation models on the KITTI benchmark dataset.
Advisors
Yoon, Kuk-Jinresearcher윤국진researcher
Description
한국과학기술원 :로봇공학학제전공,
Publisher
한국과학기술원
Issue Date
2020
Identifier
325007
Language
eng
Description

학위논문(석사) - 한국과학기술원 : 로봇공학학제전공, 2020.2,[ii, 43 p. :]

Keywords

Spatio-temporal consistency▼aLoop consistency loss▼aStereo matching▼aOptical flow▼aUnsupervised learning▼aZNCC(Zero Mean Normalized Cross-Correlation); 시공간일관성▼a루프일관성 손실함수▼a스테레오 정합▼a광학흐름 추정▼a비지도학습; ZNCC(Zero Mean Normalized Cross-Correlation)

URI
http://hdl.handle.net/10203/283984
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=910775&flag=dissertation
Appears in Collection
RE-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0