Real-time human segmentation from RGB-D video sequence based on adaptive geodesic distance computation

Yeong Seok Kim, Jong Chul Yoon, In Kwon Lee

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In this paper, we propose a method for extracting humans in the foreground of video frames using color and depth information. To ensure real-time performance and to increase accuracy, we classify a video frame into two parts by degree of noise: head region with high noise level, and non-head region with low noise level. Then, we apply a high-computational geodesic matting algorithm to the noisy head region that includes hair, and a low-computational hole filling with smoothing method to other regions. Additionally, we modify the traditional color-based geodesic segmentation algorithm to consider additional depth information. Then, we apply temporal/spatial smoothing to the blended foreground mask in order to enhance the coherence between video frames. Experimental results show that the proposed method outperforms a previous approach by accuracy and performance.

Original languageEnglish
Pages (from-to)28409-28421
Number of pages13
JournalMultimedia Tools and Applications
Volume78
Issue number20
DOIs
Publication statusPublished - 2019 Oct 1

Fingerprint

Color
Masks

All Science Journal Classification (ASJC) codes

  • Software
  • Media Technology
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

@article{393111f80b0048bab2772bf441b03641,
title = "Real-time human segmentation from RGB-D video sequence based on adaptive geodesic distance computation",
abstract = "In this paper, we propose a method for extracting humans in the foreground of video frames using color and depth information. To ensure real-time performance and to increase accuracy, we classify a video frame into two parts by degree of noise: head region with high noise level, and non-head region with low noise level. Then, we apply a high-computational geodesic matting algorithm to the noisy head region that includes hair, and a low-computational hole filling with smoothing method to other regions. Additionally, we modify the traditional color-based geodesic segmentation algorithm to consider additional depth information. Then, we apply temporal/spatial smoothing to the blended foreground mask in order to enhance the coherence between video frames. Experimental results show that the proposed method outperforms a previous approach by accuracy and performance.",
author = "Kim, {Yeong Seok} and Yoon, {Jong Chul} and Lee, {In Kwon}",
year = "2019",
month = "10",
day = "1",
doi = "10.1007/s11042-017-5375-5",
language = "English",
volume = "78",
pages = "28409--28421",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "20",

}

Real-time human segmentation from RGB-D video sequence based on adaptive geodesic distance computation. / Kim, Yeong Seok; Yoon, Jong Chul; Lee, In Kwon.

In: Multimedia Tools and Applications, Vol. 78, No. 20, 01.10.2019, p. 28409-28421.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Real-time human segmentation from RGB-D video sequence based on adaptive geodesic distance computation

AU - Kim, Yeong Seok

AU - Yoon, Jong Chul

AU - Lee, In Kwon

PY - 2019/10/1

Y1 - 2019/10/1

N2 - In this paper, we propose a method for extracting humans in the foreground of video frames using color and depth information. To ensure real-time performance and to increase accuracy, we classify a video frame into two parts by degree of noise: head region with high noise level, and non-head region with low noise level. Then, we apply a high-computational geodesic matting algorithm to the noisy head region that includes hair, and a low-computational hole filling with smoothing method to other regions. Additionally, we modify the traditional color-based geodesic segmentation algorithm to consider additional depth information. Then, we apply temporal/spatial smoothing to the blended foreground mask in order to enhance the coherence between video frames. Experimental results show that the proposed method outperforms a previous approach by accuracy and performance.

AB - In this paper, we propose a method for extracting humans in the foreground of video frames using color and depth information. To ensure real-time performance and to increase accuracy, we classify a video frame into two parts by degree of noise: head region with high noise level, and non-head region with low noise level. Then, we apply a high-computational geodesic matting algorithm to the noisy head region that includes hair, and a low-computational hole filling with smoothing method to other regions. Additionally, we modify the traditional color-based geodesic segmentation algorithm to consider additional depth information. Then, we apply temporal/spatial smoothing to the blended foreground mask in order to enhance the coherence between video frames. Experimental results show that the proposed method outperforms a previous approach by accuracy and performance.

UR - http://www.scopus.com/inward/record.url?scp=85033385657&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85033385657&partnerID=8YFLogxK

U2 - 10.1007/s11042-017-5375-5

DO - 10.1007/s11042-017-5375-5

M3 - Article

AN - SCOPUS:85033385657

VL - 78

SP - 28409

EP - 28421

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 20

ER -