Page 121 - 《渔业研究》2026年第1期

118 渔 业 研 究 第 48 卷

Individual identification of large yellow croaker (Larimichthys crocea)
based on computer vision

ZHAO Yaning，GU Linlin，YANG Zhe，JIANG Dan，WANG Zhiyong，FANG Ming *
(Key Laboratory of Healthy Mariculture for the East China Sea, Ministry of Agriculture and Rural Affairs & Fisheries college,
Jimei University, Xiamen 361021, China)

Abstract: [Background] Individual identification is crucial for feed nutrition and genetic breeding in fish
aquaculture. Passive integrated transponder (PIT) tagging is currently the mainstream method for fish individu-
al identification, but this technology has several unresolved limitations. [Objective] To address the invasive
damage, high material costs, and low efficiency associated with PIT tagging, this study aims to develop a uni-
versal visual recognition technology applicable to fish lacking distinct phenotypic features (e.g., skin spots or
stripes). Using the large yellow croaker (Larimichthys crocea), an economically important species in the East
China Sea, as the validation subject, the study systematically evaluated the cross-temporal-scale individual iden-
tification capability of this technology. [Methods] The study established a fish individual identification system
using a ResNet50 network architecture with residual structures as the backbone. The system learned discriminat-
ive features between individual images and constructed a recognition feature database. [Results] The study de-
veloped an image acquisition protocol and collected a total of 7 960 bilateral images and 1 410 dorsal-ventral im-
ages from the same batch of L. crocea during their genetic breeding phase in the Baiji Bay area, Fujian
Province, filling the gap in the current individual image database for this species. The feature learning model
was trained using images from 2 061 fish collected 8 weeks before spawning and tested for recognition 1 week
before spawning (7 weeks later). Test results showed that the proposed method achieved a short-term recogni-
tion accuracy rate of 95.20% using bilateral images, with medium and long-term accuracy rate (across test
groups) of (82.90±1.98)%. When using single-side images for medium and long-term recognition, the accuracy
rate dropped to (77.70±3.23)% (side 1) and (74.50±1.41)% (side 2), representing a 3.00%–8.50% reduction
compared to bilateral images, suggesting that bilateral image data should be prioritized in practical applications.
Additionally, models trained on dorsal-ventral images achieved a maximum validation accuracy rate of only
10.78%, confirming the superior efficacy of bilateral images for biometric feature extraction and individual
identification. [Conclusion] The individual identification technology developed in the study exhibits temporal
stability unaffected by morphological changes. It provides foundational support for genetic breeding and feed
nutrition research in L. crocea and offers novel insights and methodologies for individual identification in other
fish species.
Key words: individual identification; computer vision; feature learning; Larimichthys crocea