Grapevine Leaves Image Dataset

KOKLU Murat (a), UNLERSEN M. Fahri (b), OZKAN Ilker Ali (a), ASLAN M. Fatih(c), SABANCI Kadir (c) (a) Department of Computer Engineering, Selcuk University, Turkey, Konya, Turkey (b) Department of Electrical and Electronics Engineering, Necmettin Erbakan University, Konya, Turkey (c) Department of Electrical-Electronic Engineering, Karamanoglu Mehmetbey University, Karaman, Turkey DATASET: https://www.muratkoklu.com/datasets/ Citation Request : Koklu, M., Unlersen, M. F., Ozkan, I. A., Aslan, M. F., & Sabanci, K. (2022). A CNN-SVM study based on selected deep features for grapevine leaves classification. Measurement, 188, 110425. Doi:https://doi.org/10.1016/j.measurement.2021.110425 Link: https://doi.org/10.1016/j.measurement.2021.110425 DATASET: https://www.muratkoklu.com/datasets/ Highlights • Classification of five classes of grapevine leaves by MobileNetv2 CNN Model. • Classification of features using SVMs with different kernel functions. • Implementing a feature selection algorithm for high classification percentage. • Classification with highest accuracy using CNN-SVM Cubic model. Abstract: The main product of grapevines is grapes that are consumed fresh or processed. In addition, grapevine leaves are harvested once a year as a by-product. The species of grapevine leaves are important in terms of price and taste. In this study, deep learning-based classification is conducted by using images of grapevine leaves. For this purpose, images of 500 vine leaves belonging to 5 species were taken with a special self-illuminating system. Later, this number was increased to 2500 with data augmentation methods. The classification was conducted with a state-of-art CNN model fine-tuned MobileNetv2. As the second approach, features were extracted from pre-trained MobileNetv2′s Logits layer and classification was made using various SVM kernels. As the third approach, 1000 features extracted from MobileNetv2′s Logits layer were selected by the Chi-Squares method and reduced to 250. Then, classification was made with various SVM kernels using the selected features. The most successful method was obtained by extracting features from the Logits layer and reducing the feature with the Chi-Squares method. The most successful SVM kernel was Cubic. The classification success of the system has been determined as 97.60%. It was observed that feature selection increased the classification success although the number of features used in classification decreased. Keywords: Deep learning, Transfer learning, SVM, Grapevine leaves, Leaf identification

科克吕·穆拉特(a)、于勒尔森·M·法赫里(b)、厄兹坎·伊尔克尔·阿里(a)、阿斯兰·M·法提赫(c)、萨班哲·卡德尔(c)
(a) 土耳其科尼亚市塞尔丘克大学计算机工程系；(b) 土耳其科尼亚市内克梅廷·埃尔巴坎大学电气与电子工程系；(c) 土耳其卡拉曼市卡拉曼奥卢·穆罕默德贝伊大学电气电子工程系

数据集：https://www.muratkoklu.com/datasets/
引文要求：Koklu, M., Unlersen, M. F., Ozkan, I. A., Aslan, M. F., & Sabanci, K. (2022). 基于精选深度特征的CNN-SVM研究用于葡萄叶片分类. 《测量》(Measurement), 188, 110425. Doi:https://doi.org/10.1016/j.measurement.2021.110425 链接：https://doi.org/10.1016/j.measurement.2021.110425

数据集：https://www.muratkoklu.com/datasets/

研究亮点：
• 采用MobileNetv2卷积神经网络（Convolutional Neural Network, CNN）模型对5类葡萄叶片进行分类
• 采用支持向量机（Support Vector Machine, SVM）结合不同核函数完成特征分类
• 针对高分类准确率设计特征选择算法
• 采用CNN-SVM三次核模型实现最优分类精度

摘要：葡萄的核心产物为可鲜食或加工的葡萄果实，而葡萄叶片作为副产物每年仅采收一次。葡萄叶片的品种直接影响其价格与风味品质。本研究依托葡萄叶片图像开展基于深度学习的分类任务。研究团队首先采用专用自照明系统采集了隶属于5个品种的500张葡萄叶片图像，随后通过数据增强技术将样本规模扩充至2500张。
分类实验共设计三种方案：第一种方案采用经过微调的前沿卷积神经网络（Convolutional Neural Network, CNN）模型MobileNetv2完成分类；第二种方案为从预训练MobileNetv2的Logits层提取特征，使用多种支持向量机（Support Vector Machine, SVM）核函数开展分类；第三种方案为通过卡方检验法从MobileNetv2的Logits层提取的1000个特征中筛选出250个特征以实现降维，随后利用筛选后的特征结合多种SVM核函数完成分类。
实验结果显示，性能最优的方案为从Logits层提取特征并通过卡方检验法完成特征降维，其中表现最佳的SVM核函数为三次多项式核。本系统的分类准确率达到97.60%。研究表明，尽管分类所用的特征数量有所减少，但特征选择步骤仍有效提升了分类精度。

关键词：深度学习、迁移学习、支持向量机（Support Vector Machine, SVM）、葡萄叶片、叶片识别

搜集汇总

背景与挑战

背景概述

该数据集包含5种葡萄叶的2500张图像，通过MobileNetv2和SVM方法实现97.6%的分类准确率，适用于深度学习和计算机视觉研究。