DASC Descriptor

Abstract

    DASC Descriptor
  • Establishing dense visual correspondences between multiple images is fundamental task in many applications of computer vision and

    computational photography. However, finding a reliable correspondence in multi-modal or multi-spectral images still remains

    unsolved due to their challenging photometric and geometric variations. In this paper, we propose a novel dense matching

    descriptor, called dense adaptive self-correlation (DASC), to estimate multi-modal and multi-spectral dense correspondences. Based

    on an observation that self-similarity existing within images is robust to imaging modality variations, we define the descriptor with

    a series of an adaptive self-correlation similarity measure between patches sampled by a randomized receptive field pooling, in which a

    sampling pattern is obtained using a discriminative learning. The computational redundancy of dense descriptors is dramatically

    reduced by applying fast edge-aware filtering. We formulate a general image deformation model, and then analyze why the DASC

    works well for multi-modal and multi-spectral variations. Furthermore, in order to address geometric variations such as scale

    and rotation, we propose a geometry-invariant DASC (GI-DASC) descriptor that effectively leverages the DASC through a

    superpixel-based representation. For a quantitative evaluation of the GI-DASC, we build a novel multi-modal benchmark with ground

    truth annotation maps as varying photometric and geometric conditions. Experimental results demonstrate the outstanding

    performance of the DASC and its variant (GI-DASC) in many cases of multi-modal and multi-spectral dense correspondence.

Publication

  • DASC: Dense Adaptive Self-Correlation Descriptor for Multi-modal and Multi-spectral Correspondence

    Seungryong Kim, Dongbo Min, Bumsub Ham, Seungchul Ryu, Minh N. Do, and Kwanghoon Sohn

    IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2015

    Paper: high-res version (24MB) [pdf], low-res version (6.7MB) [pdf]

    Supplementary: [pdf]

    Poster: [pdf]

  • DASC: Robust Dense Descriptor for Multi-modal and Multi-spectral Correspondence Estimation

    Seungryong Kim, Dongbo Min, Bumsub Ham, Minh N. Do, and Kwanghoon Sohn

    IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 9, pp. 1712-1729, Sep. 2017

    Paper: [pdf]

    Supplementary: [pdf]

[Top]

Download

[Top]