Lightweight and efficient attention-based CNN models for In-field strawberry instance segmentation

Zhao, Ziang ORCID: https://orcid.org/0009-0004-6600-5581, Hicks, Yulia ORCID: https://orcid.org/0000-0002-7179-4587, Sun, Xianfang ORCID: https://orcid.org/0000-0002-6114-0766, McGuiness, Benjamin and Lim, Hin 2024. Lightweight and efficient attention-based CNN models for In-field strawberry instance segmentation. Presented at: 2024 IEEE 20th International Conference on Automation Science and Engineering, Bari, Italy, 28 August - 1 September 2024. 2024 IEEE 20th International Conference on Automation Science and Engineering. IEEE, pp. 3294-3299. 10.1109/CASE59546.2024.10711802

Preview

PDF - Accepted Post-Print Version Download (2MB) | Preview

Abstract

Accurately segmenting strawberries within real- world production settings not only helps the development of automated harvesting robots but also allows for precise calcula- tion of the number and size of strawberries, providing accurate yield information for agricultural planning and resource opti- mization. This paper proposes lightweight and efficient CNN models specifically designed for strawberry instance segmenta- tion, consisting of an efficient self-attention-based backbone, a feature pyramid network (FPN), and a decoder with an instance branch and a mask branch. The proposed models surpass the original and simplified Mask R-CNN with significant 21.46 and 22.97 AP gains respectively, with the Base backbone achieving the highest AP of 70.22. Additionally, our models demonstrate efficiency by requiring much fewer parameters (17.42M) and floating-point operations (78.3G) compared to Mask R-CNN (35.08M / 877.4G), making them suitable for deployment on devices with limited computational resources.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Published Online
Status:	Published
Schools:	Engineering Computer Science & Informatics
Publisher:	IEEE
ISBN:	9798350358520
ISSN:	2161-8070
Funders:	China Scholarship Council
Date of First Compliant Deposit:	9 August 2024
Date of Acceptance:	3 June 2024
Last Modified:	01 Nov 2024 09:20
URI:	https://orca.cardiff.ac.uk/id/eprint/171287

Actions (repository staff only)

Edit Item