EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

MsGf: A Lightweight Self-Supervised Monocular Depth Estimation Framework with Multi-Scale Feature Extraction

Xinxing Tian, Zhilin He, Yawei Zhang, Fengkai Liu, Tianhao Gu

Sensors · 16.10.2025 · 25 (20) · 6380

DOI: 10.3390/s25206380 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 16.10.2025. Verfasst von Xinxing Tian und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Automation, Qingdao University, Qingdao 266071, China. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, China Postdoctoral Science Foundation, Natural Science Foundation of Qingdao City.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Monocular depth estimation is an essential component in computer vision that enables 3D scene understanding, with critical applications in autonomous driving and augmented reality. This paper proposes a lightweight self-supervised framework from single RGB images for multi-scale feature extraction and artifact elimination in monocular depth estimation (MsGf). The proposed framework first designs a Cross-Dimensional Multi-scale Feature Extraction (CDMs) module. The CDMs module combines parallel multi-scale convolution with sequential feature convolutions to achieve multi-scale feature extraction with minimal parameters. Additionally, a Sobel Edge Perception-Guided Filtering (SEGF) module is proposed. The SEGF module uses the Sobel operator to decompose the features into horizontal direction features and vertical direction features, and then generates the filter kernel through two steps of filtering to effectively suppress artifacts and better capture structural and edge features. A large number of ablation experiments and comparative experiments on the KITTI and Make3D datasets demonstrate that the MsGf with only 0.8 M parameters can achieve better performance than the current most advanced methods.

Schlagwörter

Stichwörter der Autor:innen

edge feature extractionguided filteringlightweight modelmulti-scale feature extractionself-supervised monocular depth estimation

Autor:innen und Institutionen

  • School of Automation, Qingdao University, Qingdao 266071, China.

Förderung

  • National Natural Science Foundation of China
  • China Postdoctoral Science Foundation
  • Natural Science Foundation of Qingdao City
  • Shandong Provincial Postdoctoral Innovation Project
  • Qingdao Postdoctoral Applied Research Project
  • Joint Research Program of Systems Science, Qingdao University

Zitieren

APA

Tian, X., He, Z., Zhang, Y., Liu, F., & Gu, T. (2025). MsGf: A Lightweight Self-Supervised Monocular Depth Estimation Framework with Multi-Scale Feature Extraction. Sensors, 25(20), 6380. https://doi.org/10.3390/s25206380

Vancouver

Tian X, He Z, Zhang Y, Liu F, Gu T. MsGf: A Lightweight Self-Supervised Monocular Depth Estimation Framework with Multi-Scale Feature Extraction. Sensors. 2025;25(20):6380. doi:10.3390/s25206380

BibTeX
@article{Tian2025,
  author = {Tian, Xinxing and He, Zhilin and Zhang, Yawei and Liu, Fengkai and Gu, Tianhao},
  title = {MsGf: A Lightweight Self-Supervised Monocular Depth Estimation Framework with Multi-Scale Feature Extraction},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {20},
  pages = {6380},
  doi = {10.3390/s25206380}
}
Kostenlos starten →