Skip to Main content Skip to Navigation
Conference papers

Human Hair Segmentation In The Wild Using Deep Shape Prior

Yongzhe Yan 1 Anthony Berthelier 1 Stefan Duffner 2 Xavier Naturel 3 Christophe Garcia 2 Thierry Chateau 1
1 COMSEE - COMputers that SEE
ISPR - Image, système de perception, robotique
2 imagine - Extraction de Caractéristiques et Identification
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : Virtual human hair dying is becoming a popular Augmented Reality (AR) application in recent years. Human hair contains diverse color and texture information which can be significantly varied from case to case depending on different hair styles and environmental lighting conditions. However, the publicly available hair segmentation datasets are relatively small. As a result, hair segmentation can be easily interfered by the cluttered background in practical use. In this paper, we propose to integrate a shape prior into Fully Convolutional Neural Network (FCNN) to mitigate this issue. First, we utilize a FCNN with an Atrous Spatial Pyramid Pooling (ASPP) module [1] to find a human hair shape prior based on a specific distance transform. In the second stage, we combine the hair shape prior and the original image to form the input of a symmetric encoder-decoder FCNN to get the final hair segmentation output. Both quantitative and qualitative results show that our method achieves state-of-the-art performance on the publicly available LFW-Part and Figaro1k datasets.
Document type :
Conference papers
Complete list of metadata

Cited literature [12 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02891974
Contributor : Yongzhe Yan Connect in order to contact the contributor
Submitted on : Tuesday, July 7, 2020 - 11:33:53 AM
Last modification on : Thursday, September 9, 2021 - 2:36:02 PM
Long-term archiving on: : Friday, November 27, 2020 - 1:06:34 PM

File

23_CVPRW_hair_segmentation.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02891974, version 1

Citation

Yongzhe Yan, Anthony Berthelier, Stefan Duffner, Xavier Naturel, Christophe Garcia, et al.. Human Hair Segmentation In The Wild Using Deep Shape Prior. Conference on Computer Vision and Pattern Recognition Workshop (CVPR Workshop), Jun 2019, Long Beach, United States. ⟨hal-02891974⟩

Share

Metrics

Record views

104

Files downloads

67