Answers are grounded in this paper and cite their sources.
Respiratory rate is not only a critical physiological indicator for assessing the health status of fish and the quality of ecological environments but also a key indicator reflecting the metabolic intensity and stress state of fish. Therefore, real-time, non-invasive, and accurate monitoring of fish respiratory rate is of great significance. Existing studies on estimating fish respiratory rate predominantly rely on manual counting, traditional machine learning, physical sensors, or tracking mouth movements, while few reports focus on the analysis of gill movement—the core respiratory organ. To address this gap, this study proposes an intelligent method for accurate respiratory rate analysis in benthic sturgeons. First, this study collected nearly 6 GB of video data under natural light for the analysis of sturgeon respiratory rates. Secondly, a precise gill segmentation model named FishGillNet was developed, in which a Residual Feature Fusion Module (RFFM) was introduced to replace the Grouped Attention Gate (GAG) in the baseline model MK-UNet. Additionally, a Multi-Frequency Channel Attention (MFCA) was integrated into skip connections between the encoder and decoder to enhance the model’s ability to segment fine details. Finally, the segmented gill region was intelligently processed using a minimum bounding rectangle combined with Gaussian filtering for precise statistical and quantitative analysis of gill movement cycles. This approach effectively circumvents the high-error challenges associated with area-based statistical methods. The experimental results show that the constructed FishGillNet semantic segmentation model achieved a mean Intersection over Union (mIoU) of 97.276 %, with a model size of 3.29 MB and a frame rate of 24.533 fps, ensuring accurate segmentation of sturgeon gills and largely meeting the requirements for real-time monitoring. By integrating FishGillNet, minimum bounding rectangle, and Gaussian filtering, this respiratory rate analysis framework achieved highly accurate sturgeon respiratory rate monitoring, with an accuracy exceeding 98 %. This study focuses on the gills, the primary respiratory organ of fish, and proposes an intelligent method for precise analysis of respiratory rate in sturgeon. The method has been preliminarily validated in juvenile sturgeon, though further investigation is needed in real-world scenarios. In summary, this method provides an automated solution for respiratory rate monitoring in future fish physiology research.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: