Gaze Target Prediction with the Understanding of 3D Scenes

Leru Gao; Fengxi Sun; Yue Liu

doi:10.1007/978-981-95-4966-5_9

Gaze Target Prediction with the Understanding of 3D Scenes

Leru Gao^*
, Fengxi Sun
, Yue Liu

^*Corresponding author for this work

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

The goal of the gaze target prediction is to determine the location where the person is focusing and the probability of the gaze falls outside the image. Although prior works have addressed this task by regressing heatmaps centered on the gaze location, they typically fail to incorporate the scene's semantic information. In this work, we first generate 3D point cloud of the given image based on depth estimation and camera intrinsics. Then we combine the point cloud and estimated 3D gaze vector to generate the 3D field of view (FoV) heatmap. Scene contextual cues are finally merged to get the output heatmap. Our method achieves competitive results on the ChildPlay, GazeFollow, and VideoAttentionTarget datasets.

Original language	English
Title of host publication	Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers
Editors	Yongtian Wang, Yi Chen
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	129-143
Number of pages	15
ISBN (Print)	9789819549658
DOIs	https://doi.org/10.1007/978-981-95-4966-5_9
Publication status	Published - 2026
Externally published	Yes
Event	20th Chinese Conference on Image and Graphics Technologies and Applications, IGTA 2025 - Beijing, China Duration: 9 Aug 2025 → 10 Aug 2025

Publication series

Name	Communications in Computer and Information Science
Volume	2800 CCIS
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	20th Chinese Conference on Image and Graphics Technologies and Applications, IGTA 2025
Country/Territory	China
City	Beijing
Period	9/08/25 → 10/08/25

Keywords

3D scene understanding
gaze estimation
gaze target prediction

Access to Document

10.1007/978-981-95-4966-5_9

Cite this

Gao, L., Sun, F., & Liu, Y. (2026). Gaze Target Prediction with the Understanding of 3D Scenes. In Y. Wang, & Y. Chen (Eds.), Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers (pp. 129-143). (Communications in Computer and Information Science; Vol. 2800 CCIS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-95-4966-5_9

Gao, Leru ; Sun, Fengxi ; Liu, Yue. / Gaze Target Prediction with the Understanding of 3D Scenes. Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers. editor / Yongtian Wang ; Yi Chen. Springer Science and Business Media Deutschland GmbH, 2026. pp. 129-143 (Communications in Computer and Information Science).

@inproceedings{6f3d0e5958ce48258108da709dd703bf,

title = "Gaze Target Prediction with the Understanding of 3D Scenes",

abstract = "The goal of the gaze target prediction is to determine the location where the person is focusing and the probability of the gaze falls outside the image. Although prior works have addressed this task by regressing heatmaps centered on the gaze location, they typically fail to incorporate the scene's semantic information. In this work, we first generate 3D point cloud of the given image based on depth estimation and camera intrinsics. Then we combine the point cloud and estimated 3D gaze vector to generate the 3D field of view (FoV) heatmap. Scene contextual cues are finally merged to get the output heatmap. Our method achieves competitive results on the ChildPlay, GazeFollow, and VideoAttentionTarget datasets.",

keywords = "3D scene understanding, gaze estimation, gaze target prediction",

author = "Leru Gao and Fengxi Sun and Yue Liu",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.; 20th Chinese Conference on Image and Graphics Technologies and Applications, IGTA 2025 ; Conference date: 09-08-2025 Through 10-08-2025",

year = "2026",

doi = "10.1007/978-981-95-4966-5\_9",

language = "English",

isbn = "9789819549658",

series = "Communications in Computer and Information Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "129--143",

editor = "Yongtian Wang and Yi Chen",

booktitle = "Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers",

address = "Germany",

}

Gao, L, Sun, F & Liu, Y 2026, Gaze Target Prediction with the Understanding of 3D Scenes. in Y Wang & Y Chen (eds), Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers. Communications in Computer and Information Science, vol. 2800 CCIS, Springer Science and Business Media Deutschland GmbH, pp. 129-143, 20th Chinese Conference on Image and Graphics Technologies and Applications, IGTA 2025, Beijing, China, 9/08/25. https://doi.org/10.1007/978-981-95-4966-5_9

Gaze Target Prediction with the Understanding of 3D Scenes. / Gao, Leru; Sun, Fengxi; Liu, Yue.
Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers. ed. / Yongtian Wang; Yi Chen. Springer Science and Business Media Deutschland GmbH, 2026. p. 129-143 (Communications in Computer and Information Science; Vol. 2800 CCIS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Gaze Target Prediction with the Understanding of 3D Scenes

AU - Gao, Leru

AU - Sun, Fengxi

AU - Liu, Yue

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.

PY - 2026

Y1 - 2026

N2 - The goal of the gaze target prediction is to determine the location where the person is focusing and the probability of the gaze falls outside the image. Although prior works have addressed this task by regressing heatmaps centered on the gaze location, they typically fail to incorporate the scene's semantic information. In this work, we first generate 3D point cloud of the given image based on depth estimation and camera intrinsics. Then we combine the point cloud and estimated 3D gaze vector to generate the 3D field of view (FoV) heatmap. Scene contextual cues are finally merged to get the output heatmap. Our method achieves competitive results on the ChildPlay, GazeFollow, and VideoAttentionTarget datasets.

AB - The goal of the gaze target prediction is to determine the location where the person is focusing and the probability of the gaze falls outside the image. Although prior works have addressed this task by regressing heatmaps centered on the gaze location, they typically fail to incorporate the scene's semantic information. In this work, we first generate 3D point cloud of the given image based on depth estimation and camera intrinsics. Then we combine the point cloud and estimated 3D gaze vector to generate the 3D field of view (FoV) heatmap. Scene contextual cues are finally merged to get the output heatmap. Our method achieves competitive results on the ChildPlay, GazeFollow, and VideoAttentionTarget datasets.

KW - 3D scene understanding

KW - gaze estimation

KW - gaze target prediction

UR - https://www.scopus.com/pages/publications/105027932979

U2 - 10.1007/978-981-95-4966-5_9

DO - 10.1007/978-981-95-4966-5_9

M3 - Conference contribution

AN - SCOPUS:105027932979

SN - 9789819549658

T3 - Communications in Computer and Information Science

SP - 129

EP - 143

BT - Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers

A2 - Wang, Yongtian

A2 - Chen, Yi

PB - Springer Science and Business Media Deutschland GmbH

T2 - 20th Chinese Conference on Image and Graphics Technologies and Applications, IGTA 2025

Y2 - 9 August 2025 through 10 August 2025

ER -

Gao L, Sun F, Liu Y. Gaze Target Prediction with the Understanding of 3D Scenes. In Wang Y, Chen Y, editors, Image and Graphics Technologies and Applications - 20th Chinese Conference, IGTA 2025, Revised Selected Papers. Springer Science and Business Media Deutschland GmbH. 2026. p. 129-143. (Communications in Computer and Information Science). doi: 10.1007/978-981-95-4966-5_9

Gaze Target Prediction with the Understanding of 3D Scenes

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this