用深度学习自动识别土耳其五种本地橄榄品种，高效模型表现更优。
Image-Based Classification of Olive Varieties Native to Turkiye Using Multiple Deep Learning Architectures: Analysis of Performance, Complexity, and Generalization
- 对比10种模型在2500张图像上分类，采用迁移学习训练。
- EfficientNetV2-S准确率达95.8%最高，EfficientNetB0综合性能最佳。
- 数据有限时，模型参数效率比深度更重要，适合资源受限场景。
本研究比较了多种深度学习架构在土耳其本土栽培的五种黑橄榄品种（Gemlik、Ayvalik、Uslu、Erkence、Celebi）图像分类中的表现。基于2500张图像的数据集，使用迁移学习训练了10种模型：MobileNetV2、EfficientNetB0、EfficientNetV2-S、ResNet50、ResNet101、DenseNet121、InceptionV3、ConvNeXt-Tiny、ViT-B16和Swin-T。通过准确率、精确率、召回率、F1分数、马修斯相关系数（MCC）、Cohen's Kappa、ROC-AUC、参数量、浮点运算数（FLOPs）、推理时间及泛化差距等指标评估性能。结果表明，EfficientNetV2-S达到最高分类准确率（95.8%），而EfficientNetB0在准确率与计算复杂度之间实现最佳平衡。整体显示，在数据有限条件下，参数效率比模型深度本身更具决定性作用。
原文摘要 · Abstract (English)
This study compares multiple deep learning architectures for the automated, image-based classification of five locally cultivated black table olive varieties in Turkey: Gemlik, Ayvalik, Uslu, Erkence, and Celebi. Using a dataset of 2500 images, ten architectures - MobileNetV2, EfficientNetB0, EfficientNetV2-S, ResNet50, ResNet101, DenseNet121, InceptionV3, ConvNeXt-Tiny, ViT-B16, and Swin-T - were trained using transfer learning. Model performance was evaluated using accuracy, precision, recall, F1-score, Matthews Correlation Coefficient (MCC), Cohen's Kappa, ROC-AUC, number of parameters, FLOPs, inference time, and generalization gap. EfficientNetV2-S achieved the highest classification accuracy (95.8%), while EfficientNetB0 provided the best trade-off between accuracy and computational complexity. Overall, the results indicate that under limited data conditions, parametric efficiency plays a more critical role than model depth alone.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。