DooDLeNet: Double DeepLab Enhanced Feature Fusion for Thermal-Color Semantic Segmentation

Oriel Frigo, Lucien Martin-Gaffé, Catherine Wacongne

CVPRW 2022 pp. 3020-3028

doi:10.1109/CVPRW56347.2022.00341 /cvprw/2022/frigo2022cvprw-doodlenet/

Abstract

In this paper we present a new approach for feature fusion between RGB and LWIR Thermal images for the task of semantic segmentation for driving perception. We propose DooDLeNet, a double DeepLab architecture with specialized encoder-decoders for thermal and color modalities and a shared decoder for final segmentation. We combine two strategies for feature fusion: confidence weighting and correlation weighting. We report state-of-the-art mean IoU results on the MF dataset [1].

PDF CVPRW Semantic Scholar

Cite

Text

Frigo et al. "DooDLeNet: Double DeepLab Enhanced Feature Fusion for Thermal-Color Semantic Segmentation." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022. doi:10.1109/CVPRW56347.2022.00341

Markdown

[Frigo et al. "DooDLeNet: Double DeepLab Enhanced Feature Fusion for Thermal-Color Semantic Segmentation." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022.](https://mlanthology.org/cvprw/2022/frigo2022cvprw-doodlenet/) doi:10.1109/CVPRW56347.2022.00341

BibTeX

@inproceedings{frigo2022cvprw-doodlenet,
  title     = {{DooDLeNet: Double DeepLab Enhanced Feature Fusion for Thermal-Color Semantic Segmentation}},
  author    = {Frigo, Oriel and Martin-Gaffé, Lucien and Wacongne, Catherine},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2022},
  pages     = {3020-3028},
  doi       = {10.1109/CVPRW56347.2022.00341},
  url       = {https://mlanthology.org/cvprw/2022/frigo2022cvprw-doodlenet/}
}