Earticle

초록

영어

Animal pattern recognition from nighttime grayscale images is crucial for wildlife protection and ecological monitoring. Most of the current models suffer from a large parameter scale, making them unsuitable for deployment in resource-constrained environments. To address this challenge, this study proposes a multi-layer knowledge distillation approach based on the teacher Convnext model to improve the lightweight student model's classification performance effectively. Experimental results show that the parameters of the distilled CifarResNet20 model are only 0.27M, and the accuracy is 88.76%, which is superior to the traditional single-layer distillation and another tiny student model. The study confirms the efficiency and practical value of the proposed method in practical applications such as ecological monitoring.

목차

ABSTRACT
1. Introduction
2. Related Work
3. Materials and Methods
3.1 Datasets
3.2 ConvNeXt Model
3.3 Distillation scheme design
4. Result
4.1 Comparison of results of different distillation schemes
4.2 Accuracy vs. Parameters
5. Conclusion
참고문헌

키워드

저자

• Dae-Won Park [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Changyu AO [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ] Corresponding Author
• Seung-Eon Jeong [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Soo-Kyung Moon [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Youn-Mo Soung [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Man-Sung Kwen [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Uk Cho [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Dae-In Kang [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Sung-Ho Jung [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]
• Gwang-Jun Kim [ Department of Computer Engineering, Chonnam National University, Yeosu, Seoul, Korea ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  컴퓨터게임및콘텐츠논문지(구 한국컴퓨터게임학회논문지) [Journal of Computer Games and Contents]

• 간기

  월간

• pISSN

  3091-7409

• eISSN

  3092-3638

• 수록기간

  2002~2026

• 등재여부

  KCI 등재

• 십진분류

  KDC 691 DDC 793