Cross Dataset Evaluation of Feature Extraction Techniques for Leaf Classification

Reul, Christian; Toepfer, Martin; Puppe, Frank

doi:10.5121/ijaia.2016.7201

Cited by 3 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, our main goal is to apply the described methods to cross dataset validation. Comparable to [18] we want to study how a trained CNN applies to a real world scenario, i.e. classifying leaves that were collected completely independent of the test set and therefore differ quite a lot due influences of the temperature, rainfall, solar irradiation or seasons.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Leaf Identification Using a Deep Convolutional Neural Network

Wick,

Puppe

2017

Preprint

Self Cite

View full text Add to dashboard Cite

Convolutional neural networks (CNNs) have become popular especially in computer vision in the last few years because they achieved outstanding performance on different tasks, such as image classifications. We propose a ninelayer CNN for leaf identification using the famous Flavia and Foliage datasets. Usually the supervised learning of deep CNNs requires huge datasets for training. However, the used datasets contain only a few examples per plant species. Therefore, we apply data augmentation and transfer learning to prevent our network from overfitting. The trained CNNs achieve recognition rates above 99% on the Flavia and Foliage datasets, and slightly outperform current methods for leaf classification.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Reul et al [18] use a 1-nearest-neighbor classifier based on contour, curvature, color, Hu, HOCS, and binary pattern features. They achieve accuracies of about 99.37% on the Flavia dataset and 95.83% on the Foliage dataset.…”

Section: Introductionmentioning

confidence: 99%

Leaf Identification Using a Deep Convolutional Neural Network

Wick,

Puppe

2017

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…It also removes holes in leaf images, extracts curvatures from boundaries, and measures smooth as well as serrated margins. This technique was used in the paper [39] to extract the arc and area features of lobeshaped leaf margins, but it is not suitable for all leaves. This technique was also used for Costa Rican species as well in [40].…”

Section: Point and Edge-based Feature Descriptorsmentioning

confidence: 99%

“…Hu, shape, texture 100% [151] Back propagation neural network Shapes, Angles and sinus of leaves 111 leaves with 14 species [125] Texture and wavelet feature Grape varieties 93.3% [126] Colour Affected and Unaffected Vegetables [127] Texture of Colour Co-occurrence Disease Identification [128] Edge Features of leaf Neem, pine oak 90.33% [129] Leaf Margin Jasmine, arka, mango, neem and shigru 85% [130] Morphological features 450 leaves of 16 classes in Ayurveda and agriculture 90% [62] Texture Foxtail, crabgrass, velvet leaf, morning glory 97% [177] Fuzzy based classifiers Statistical features of leaf 97.6% [26] Texture, shape, colour 99.87% [86] K-nearest neighbour Edge, vein, ring projection wavelet feature 87.14% [120] Geometrical features 80% [134] Leaf Margin? texture 75.5% [135] Texton Costa Rican Flavia Dataset 99.1% [40] Texton 87.14% [120] Texture ICL-97.07% Plumber-72.8% Simthsonain-73.08% [37] HoCS, contour, colour, curvature Flavia 99.61% [39] Texture 97.55% [94] Run length sequence 93.17% [152] Contour-amplitude frequency descriptor Swedish-89.6% ICL-91.6% [198] Moving centre classifier Moment invariant 92.6% [137] Bayesian classifier Fourier descriptor 88% [116] Support vector machine HoCS Leafsnap [38] Wavelet features Ornamental Plants 95.83% [114] Fourier and texture Australian Federal dataset-100%, Flavia-99.7%, Foliage-99.8%, Swedish and Middle European datasets-99.2% [93] Kernel level descriptor Flavia-97.5% [110,111] Hu moments Annona Squamosa and Psidiuguajava, 86.6% [139] Lanculariity Flavia-95.048% [84] HOG? Zernike moments Sweedish-97 [192] Map reduce algorithm Texture Hierarchical big database-91% [196] 7.3 The ICL Dataset…”

Section: The Flavia Datasetmentioning

confidence: 99%

“…They extracted features such as image moments, Fourier descriptors and leaf size and classified using the KNN classifier and achieved 88.9% accuracy. The authors of the papers [39] used the Histogram Over Curvature scale feature with the 1 Nearest neighbour classifier, obtaining an accuracy of 95.66%. Cerutti et al [184] used 1000 compound leaf images of 17 European tree species.…”

Section: The Middle European Wood Databasementioning

confidence: 99%

See 1 more Smart Citation

A Review of Visual Descriptors and Classification Techniques Used in Leaf Species Identification

Thyagharajan¹,

Raji²

2018

Arch Computat Methods Eng

View full text Add to dashboard Cite

Plants are fundamentally important to life. Key research areas in plant science include plant species identification, weed classification using hyper spectral images, monitoring plant health and tracing leaf growth, and the semantic interpretation of leaf information. Botanists easily identify plant species by discriminating between the shape of the leaf, tip, base, leaf margin and leaf vein, as well as the texture of the leaf and the arrangement of leaflets of compound leaves. Because of the increasing demand for experts and calls for biodiversity, there is a need for intelligent systems that recognize and characterize leaves so as to scrutinize a particular species, the diseases that affect them, the pattern of leaf growth, and so on. We review several image processing methods in the feature extraction of leaves, given that feature extraction is a crucial technique in computer vision. As computers cannot comprehend images, they are required to be converted into features by individually analysing image shapes, colours, textures and moments. Images that look the same may deviate in terms of geometric and photometric variations. In our study, we also discuss certain machine learning classifiers for an analysis of different species of leaves. Table 2 Inferences of data sources Visual descriptors of leaf and classification No. of papers selected Geometrical descriptors 30 Leaf shape/tip/base/venation 40 Texture/texton 25

show abstract

Plant Identification in a Combined-Imbalanced Leaf Dataset

2022

View full text Add to dashboard Cite

Plant identification has applications in ethnopharmacology and agriculture. Since leaves are one of a distinguishable feature of a plant, they are routinely used for identification. Recent developments in deep learning have made it possible to accurately identify the majority of samples in five publicly available leaf datasets. However, each dataset captures the images in a highly controlled environment. This paper evaluates the performance of EfficientNet and several other convolutional neural network (CNN) architectures when applied to a combination of the LeafSnap, Middle European Woody Plants 2014, Flavia, Swedish, and Folio datasets. To normalize the impact of imbalance resulting from combining the original datasets, we used oversampling, undersampling, and transfer learning techniques to construct an end-to-end CNN classifier. We placed greater emphasis on metrics appropriate for a diverse-imbalanced dataset rather than stressing high performance on any one of the original datasets. A model from EfficientNet's family of CNN models achieved a highly accurate F-score of 0.9861 on the combined dataset.INDEX TERMS Leaf dataset, imbalanced dataset, convolutional neural networks, transfer learning, plant identification.

show abstract

Cross Dataset Evaluation of Feature Extraction Techniques for Leaf Classification

Abstract: ABSTRACT

Cited by 3 publications

References 15 publications

Leaf Identification Using a Deep Convolutional Neural Network

Leaf Identification Using a Deep Convolutional Neural Network

A Review of Visual Descriptors and Classification Techniques Used in Leaf Species Identification

Plant Identification in a Combined-Imbalanced Leaf Dataset

Contact Info

Product

Resources

About