Lightweighting Method for DeepLabv3+ Based on MobileNet 


Vol. 14,  No. 11, pp. 919-924, Nov.  2025
https://doi.org/10.3745/TKIPS.2025.14.11.919


PDF
  Abstract

Recent advances in autonomous driving, medical image analysis, and surveillance technologies have led to an increasing demand for real-time image segmentation. Consequently, research on real-time segmentation methods has been actively pursued, with deep learning-based approaches demonstrating high performance metrics. However, the complex architectures, massive computational cost, and large parameter counts inherent in artificial neural networks impose significant constraints when deploying them on resource-limited embedded platforms. To address this issue, various attempts have been made, such as employing lightweight classification networks like MobileNet as backbone networks for segmentation models. Nevertheless, segmentation networks still involve substantial computational overhead and parameter size, making them difficult to apply in resource-constrained environments. Therefore, even when lightweight classification networks are used, further compression is necessary. In this paper, we propose a lightweight approach for DeepLabv3+ which MobileNetV3-Large without the last expansion layer as the backbone network. To demonstrate the effectiveness of the proposed method, we compare it with several DeepLabv3+ architectures whose backbone networks are MobileNetV3-Large, MobileNetV3-Small, MobileNetV3-Small without the last expansion layer, or MobileNetV3-Large compressed by hyperparameter tuning. Compared to the best performing architecture with MobileNetV3-Large as the backbone network, the structure with the proposed method can reduce the parameters by about 52% with only a 2.7%p performance decrease. Therefore, the proposed method effectively achieves the trade-off between performance and neural network size, which confirms that it can be practically utilized in resource-constrained environments.

  Statistics
  Cite this article

[IEEE Style]

T. Kim, I. Jeong, J. Kang, S. Jo, B. Moon, "Lightweighting Method for DeepLabv3+ Based on MobileNet," The Transactions of the Korea Information Processing Society, vol. 14, no. 11, pp. 919-924, 2025. DOI: https://doi.org/10.3745/TKIPS.2025.14.11.919.

[ACM Style]

Taejun Kim, Insu Jeong, Joowan Kang, Seungjun Jo, and Byungin Moon. 2025. Lightweighting Method for DeepLabv3+ Based on MobileNet. The Transactions of the Korea Information Processing Society, 14, 11, (2025), 919-924. DOI: https://doi.org/10.3745/TKIPS.2025.14.11.919.

Forms

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)

Old Journals

Indexing

All publications of TKIPS is indexed in DOI, EBSCO, Google Scholar, Crossref, and CrossCheck.

TKIPS is also selected as the Journal for Accreditation by NRF (National Research Foundation of Korea).

Related Journals