基于SE-ResNet18模型的三疣梭子蟹性别分类方法

Research on sex classification method of Portunus trituberculatus based on SE-ResNet18 model

三疣梭子蟹(Portunus trituberculatus)是一种具有重要经济价值的甲壳类动物,性别分类的准确性对于优化水产养殖策略和提高海产品加工效率至关重要。该研究提出了一种基于深度学习的自动性别分类方法,采用了增强型卷积神经网络模型SE-ResNet18。该模型结合了Squeeze-and-Excitation(SE)模块和全局平均池化,通过数据增强和优化算法对包含大量雌雄梭子蟹图像的数据集进行了训练和验证。结果显示,SE-ResNet18的总体分类准确率达到99.5%,相比ResNet18提高了近4个百分点,其中雄性梭子蟹的分类准确率为99.68%,雌性梭子蟹为99.74%。研究表明,SE-ResNet18在三疣梭子蟹的性别分类任务中具有极高的准确性和鲁棒性,能够高效地完成自动化分类任务。

Portunus trituberculatus,a crustacean species of significant economic value,is important in aquaculture and seafood processing.Accurate sex classification is critical for optimizing aquaculture strategies and improving operational efficiency in these industries.Traditional manual methods for sex classification are labor-intensive and prone to errors,highlighting the need for automated solutions.This study proposes an automated sex classification method based on deep learning,utilizing the SE-ResNet18 model,an enhanced variant of ResNet-18.The SE-ResNet18 model incorporates the Squeeze-and-Excitation(SE)module and global average pooling,enabling it to emphasize key feature channels selectively.The model was trained and validated on a large dataset of male and female Portunus trituberculatus images,with data augmentation techniques applied to improve generalization.The results show that SE-ResNet18 achieves a classification accuracy of 99.5%,nearly 4 percentage points higher than ResNet-18.Specifically,male crabs were classified with 99.68%accuracy,and female crabs with 99.74%.The model's robustness was tested under varying conditions,confirming its suitability for real-world applications in automated seafood processing and aquaculture.In conclusion,SE-ResNet18 offers a highly accurate and scalable solution for sex classification in Portunus trituberculatus,with the potential to significantly enhance productivity and efficiency in the aquaculture industry.