发表机构
Vietnamese-German University（越德大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究首次将ViT和自监督学习应用于虾病检测，提出监督迁移学习与SimCLR对比学习两种流水线，在4,348张图像上分别实现96%和85%的准确率，为水产养殖监测建立新基线。
AI 中文摘要
虾类疾病持续给水产养殖业造成毁灭性损失，推动了对稳健、自动化检测的迫切需求。本研究首次将视觉变换器（ViT）和自监督学习（SSL）应用于养虾领域，解决了性能瓶颈和数据标注挑战。我们提出了两种深度学习流水线，用于对四种关键疾病进行分类：健康、黑鳃病（BG）、白斑综合征病毒（WSSV）以及两者的共感染，使用包含4,348张图像的数据集。首先，我们的监督迁移学习方法利用了ImageNet预训练的ViT-Small/16和EfficientNet骨干网络。其次，我们引入了一个对比学习框架（SimCLR），使用ViT-Small编码器在微调之前从无标签图像中提取稳健表示。我们的结果为可持续水产养殖监测建立了强有力的新基线。监督方法实现了卓越的96%准确率和快速收敛，优于传统通用模型，而标签高效的SSL方法达到了极具竞争力的85%验证准确率。
英文摘要
Shrimp diseases continue to cause devastating losses in the aquaculture industry, driving a critical need for robust, automated detection. This work contributes the first application of Vision Transformers (ViT) and Self-Supervised Learning (SSL) to the shrimp farming domain, addressing both performance bottlenecks and data labeling challenges. We propose two deep learning pipelines to classify four key diseases: Healthy, Black Gill (BG), White Spot Syndrome Virus (WSSV), and a co-infection of both using a dataset of 4,348 images. First, our supervised transfer-learning approach leverages ImageNet-pretrained ViT-Small/16 and EfficientNet backbones. Second, we introduce a contrastive learning framework (SimCLR) with a ViT-Small encoder to extract robust representations from unlabeled images prior to fine-tuning. Our results establish strong new baselines for sustainable aquaculture monitoring. The supervised approach achieves an outstanding 96% accuracy with fast convergence, outperforming traditional generic models, while the label-efficient SSL approach reaches a highly competitive 85% validation accuracy.
CommentsThis paper has been accepted at the International Conference on Multidisciplinary Research (ICMR 2025)
