TY - GEN
T1 - Matching-space Stereo Networks for Cross-domain Generalization
AU - Cai, Changjiang
AU - Poggi, Matteo
AU - Mattoccia, Stefano
AU - Mordohai, Philippos
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/11
Y1 - 2020/11
N2 - End-to-end deep networks represent the state of the art for stereo matching. While excelling on images framing environments similar to the training set, major drops in accuracy occur in unseen domains (e.g., when moving from synthetic to real scenes). In this paper we introduce a novel family of architectures, namely Matching-Space Networks (MS-Nets), with improved generalization properties. By replacing learning-based feature extraction from image RGB values with matching functions and confidence measures from conventional wisdom, we move the learning process from the color space to the Matching Space, avoiding over-specialization to domain specific features. Extensive experimental results on four real datasets highlight that our proposal leads to superior generalization to unseen environments over conventional deep architectures, keeping accuracy on the source domain almost unaltered. Our code is available at https://qithub.com/ccj5351/MS-Nets.
AB - End-to-end deep networks represent the state of the art for stereo matching. While excelling on images framing environments similar to the training set, major drops in accuracy occur in unseen domains (e.g., when moving from synthetic to real scenes). In this paper we introduce a novel family of architectures, namely Matching-Space Networks (MS-Nets), with improved generalization properties. By replacing learning-based feature extraction from image RGB values with matching functions and confidence measures from conventional wisdom, we move the learning process from the color space to the Matching Space, avoiding over-specialization to domain specific features. Extensive experimental results on four real datasets highlight that our proposal leads to superior generalization to unseen environments over conventional deep architectures, keeping accuracy on the source domain almost unaltered. Our code is available at https://qithub.com/ccj5351/MS-Nets.
UR - http://www.scopus.com/inward/record.url?scp=85101465047&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85101465047&partnerID=8YFLogxK
U2 - 10.1109/3DV50981.2020.00046
DO - 10.1109/3DV50981.2020.00046
M3 - Conference contribution
AN - SCOPUS:85101465047
T3 - Proceedings - 2020 International Conference on 3D Vision, 3DV 2020
SP - 364
EP - 373
BT - Proceedings - 2020 International Conference on 3D Vision, 3DV 2020
T2 - 8th International Conference on 3D Vision, 3DV 2020
Y2 - 25 November 2020 through 28 November 2020
ER -