Transfer Learning for Crop classification with Cropland Data Layer data (CDL) as training samples

Due to the increasing demand for the monitoring of crop conditions and food production, it is a challenging and meaningful task to identify crops from remote sensing images. The state-of the-art crop classification models are mostly built on supervised classification models such as support vector machines (SVM), convolutional neural networks (CNN), and long- and short-term memory neural networks (LSTM). Meanwhile, as an unsupervised generative model, the adversarial generative network (GAN) is rarely used to complete classification tasks for agricultural applications. In this work, we propose a new method that combines GAN, CNN, and LSTM models to classify crops of corn and soybeans from remote sensing time-series images, in which GAN’s discriminator was used as the final classifier. The method is feasible on the condition that the training samples are small, and it fully takes advantage of spectral, spatial, and phenology features of crops from satellite data. The classification experiments were conducted on crops of corn, soybeans, and others. To verify the effectiveness of the proposed method, comparisons with models of SVM, SegNet, CNN, LSTM, and different combinations were also conducted. The results show that our method achieved the best classification results, with the Kappa coefficient of 0.7933 and overall accuracy of 0.86. Experiments in other study areas also demonstrate the extensibility of the proposed method.

Download Full-text

Robust Finger Vein Recognition Based on Deep CNN with Spatial Attention and Bias Field Correction

International Journal of Artificial Intelligence Tools ◽

10.1142/s0218213021400054 ◽

2021 ◽

Vol 30 (01) ◽

pp. 2140005

Author(s):

Zhe Huang ◽

Chengan Guo

Keyword(s):

Spatial Attention ◽

Transfer Learning ◽

Learning Strategy ◽

Bias Field ◽

The Public ◽

Finger Vein ◽

Bias Field Correction ◽

Training Samples ◽

Vein Recognition ◽

Finger Vein Recognition

As one of the biometric information based authentication technologies, finger vein recognition has received increasing attention due to its safety and convenience. However, it is still a challenging task to design an efficient and robust finger vein recognition system because of the low quality of the finger vein images, lack of sufficient number of training samples with image-level annotated information and no pixel-level finger vein texture labels in the public available finger vein databases. In this paper, we propose a novel CNN-based finger vein recognition approach with bias field correction, spatial attention mechanism and a multistage transfer learning strategy to cope with the difficulties mentioned above. In the proposed method, the bias field correction module is to remove the unbalanced bias field of the original images by using a two-dimensional polynomial fitting algorithm, the spatial attention module is to enhance the informative vein texture regions while suppressing the other less informative regions, and the multistage transfer learning strategy is to solve the problem caused by insufficient training for CNN-based model due to lack of labeled training samples in the public finger vein databases. Moreover, several measures, including a label smoothing scheme and data augmentation, are exploited to improve the performance of the proposed method. Extensive experiments have been conducted in the work on three public databases, and the results show that the proposed approach outperforms the existing state-of-the-art methods.

Download Full-text

Multisignal VGG19 Network with Transposed Convolution for Rotating Machinery Fault Diagnosis Based on Deep Transfer Learning

Shock and Vibration ◽

10.1155/2020/8863388 ◽

2020 ◽

Vol 2020 ◽

pp. 1-12

Author(s):

Jianye Zhou ◽

Xinyu Yang ◽

Lin Zhang ◽

Siyu Shao ◽

Gangying Bian

Keyword(s):

Fault Diagnosis ◽

Transfer Learning ◽

High Efficiency ◽

Image Resolution ◽

Fine Tuning ◽

Test Accuracy ◽

Multiple Sensors ◽

Learning Framework ◽

Training Samples ◽

Model Training

To realize high-precision and high-efficiency machine fault diagnosis, a novel deep learning framework that combines transfer learning and transposed convolution is proposed. Compared with existing methods, this method has faster training speed, fewer training samples per time, and higher accuracy. First, the raw data collected by multiple sensors are combined into a graph and normalized to facilitate model training. Next, the transposed convolution is utilized to expand the image resolution, and then the images are treated as the input of the transfer learning model for training and fine-tuning. The proposed method adopts 512 time series to conduct experiments on two main mechanical datasets of bearings and gears in the variable-speed gearbox, which verifies the effectiveness and versatility of the method. We have obtained advanced results on both datasets of the gearbox dataset. The dataset shows that the test accuracy is 99.99%, achieving a significant improvement from 98.07% to 99.99%.

Download Full-text

Domain Adversarial Neural Networks for Large-Scale Land Cover Classification

Remote Sensing ◽

10.3390/rs11101153 ◽

2019 ◽

Vol 11 (10) ◽

pp. 1153 ◽

Cited By ~ 1

Author(s):

Mesay Belete Bejiga ◽

Farid Melgani ◽

Pietro Beraldini

Keyword(s):

Neural Networks ◽

Land Cover ◽

Transfer Learning ◽

Large Scale ◽

Domain Adaptation ◽

Land Cover Classification ◽

Target Domain ◽

Training Samples ◽

Latent Space ◽

Large Scale Land

Learning classification models require sufficiently labeled training samples, however, collecting labeled samples for every new problem is time-consuming and costly. An alternative approach is to transfer knowledge from one problem to another, which is called transfer learning. Domain adaptation (DA) is a type of transfer learning that aims to find a new latent space where the domain discrepancy between the source and the target domain is negligible. In this work, we propose an unsupervised DA technique called domain adversarial neural networks (DANNs), composed of a feature extractor, a class predictor, and domain classifier blocks, for large-scale land cover classification. Contrary to the traditional methods that perform representation and classifier learning in separate stages, DANNs combine them into a single stage, thereby learning a new representation of the input data that is both domain-invariant and discriminative. Once trained, the classifier of a DANN can be used to predict both source and target domain labels. Additionally, we also modify the domain classifier of a DANN to evaluate its suitability for multi-target domain adaptation problems. Experimental results obtained for both single and multiple target DA problems show that the proposed method provides a performance gain of up to 40%.

Download Full-text

Effects of Class Purity of Training Patch on Classification Performance of Crop Classification with Convolutional Neural Network

Applied Sciences ◽

10.3390/app10113773 ◽

2020 ◽

Vol 10 (11) ◽

pp. 3773

Author(s):

Soyeon Park ◽

No-Wook Park

Keyword(s):

Classification Accuracy ◽

Supervised Classification ◽

Patch Size ◽

Classification Performance ◽

Spatial Homogeneity ◽

High Class ◽

Large Size ◽

Training Samples ◽

Compositional Homogeneity ◽

Crop Classification

As the performance of supervised classification using convolutional neural networks (CNNs) are affected significantly by training patches, it is necessary to analyze the effects of the information content of training patches in patch-based classification. The objective of this study is to quantitatively investigate the effects of class purity of a training patch on performance of crop classification. Here, class purity that refers to a degree of compositional homogeneity of classes within a training patch is considered as a primary factor for the quantification of information conveyed by training patches. New quantitative indices for class homogeneity and variations of local class homogeneity over the study area are presented to characterize the spatial homogeneity of the study area. Crop classification using 2D-CNN was conducted in two regions (Anbandegi in Korea and Illinois in United States) with distinctive spatial distributions of crops and class homogeneity over the area to highlight the effect of class purity of a training patch. In the Anbandegi region with high class homogeneity, superior classification accuracy was obtained when using large size training patches with high class purity (7.1%p improvement in overall accuracy over classification with the smallest patch size and the lowest class purity). Training patches with high class purity could yield a better identification of homogenous crop parcels. In contrast, using small size training patches with low class purity yielded the highest classification accuracy in the Illinois region with low class homogeneity (19.8%p improvement in overall accuracy over classification with the largest patch size and the highest class purity). Training patches with low class purity could provide useful information for the identification of diverse crop parcels. The results indicate that training samples in patch-based classification should be selected based on the class purity that reflects the local class homogeneity of the study area.

Download Full-text

Data partnership synergy: The Cropland Data Layer

2009 17th International Conference on Geoinformatics ◽

10.1109/geoinformatics.2009.5293489 ◽

2009 ◽

Cited By ~ 2

Author(s):

R. Mueller ◽

C. Boryan ◽

R. Seffrin

Keyword(s):

Data Layer ◽

Partnership Synergy ◽

Cropland Data Layer

Download Full-text

Towards CNN Representations for Small Mass Spectrometry Data Classification: From Transfer Learning to Cumulative Learning

10.1101/2020.03.24.005975 ◽

2020 ◽

Cited By ~ 1

Author(s):

Khawla Seddiki ◽

Philippe Saudemont ◽

Frédéric Precioso ◽

Nina Ogrinc ◽

Maxence Wisztorski ◽

...

Keyword(s):

Mass Spectrometry ◽

Learning Strategies ◽

Transfer Learning ◽

Classification Accuracy ◽

Data Classification ◽

Representation Learning ◽

Mass Spectrometry Data ◽

Training Dataset ◽

Require Time ◽

Training Samples

AbstractRapid and accurate clinical diagnosis of pathological conditions remains highly challenging. A very important component of diagnosis tool development is the design of effective classification models with Mass spectrometry (MS) data. Some popular Machine Learning (ML) approaches have been investigated for this purpose but these ML models require time-consuming preprocessing steps such as baseline correction, denoising, and spectrum alignment to remove non-sample-related data artifacts. They also depend on the tedious extraction of handcrafted features, making them unsuitable for rapid analysis. Convolutional Neural Networks (CNNs) have been found to perform well under such circumstances since they can learn efficient representations from raw data without the need for costly preprocessing. However, their effectiveness drastically decreases when the number of available training samples is small, which is a common situation in medical applications. Transfer learning strategies extend an accurate representation model learnt usually on a large dataset containing many categories, to a smaller dataset with far fewer categories. In this study, we first investigate transfer learning on a 1D-CNN we have designed to classify MS data, then we develop a new representation learning method when transfer learning is not powerful enough, as in cases of low-resolution or data heterogeneity. What we propose is to train the same model through several classification tasks over various small datasets in order to accumulate generic knowledge of what MS data are, in the resulting representation. By using rat brain data as the initial training dataset, a representation learning approach can have a classification accuracy exceeding 98% for canine sarcoma cancer cells, human ovarian cancer serums, and pathogenic microorganism biotypes in 1D clinical datasets. We show for the first time the use of cumulative representation learning using datasets generated in different biological contexts, on different organisms, in different mass ranges, with different MS ionization sources, and acquired by different instruments at different resolutions. Our approach thus proposes a promising strategy for improving MS data classification accuracy when only small numbers of samples are available as a prospective cohort. The principles demonstrated in this work could even be beneficial to other domains (astronomy, archaeology…) where training samples are scarce.

Download Full-text

STRATEGIC OPTIMIZATION OF CONVOLUTIONAL NEURAL NETWORKS FOR HYPERSPECTRAL LAND COVER CLASSIFICATION

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xliii-b3-2020-363-2020 ◽

2020 ◽

Vol XLIII-B3-2020 ◽

pp. 363-369

Author(s):

C. Buehler ◽

F. Schenkel ◽

W. Gross ◽

G. Schaab ◽

W. Middelmann

Keyword(s):

Transfer Learning ◽

Hyperspectral Data ◽

Electromagnetic Spectrum ◽

Training Samples ◽

Random Weights ◽

Wide Range ◽

Mixed Pixels ◽

Earth Observation Satellites ◽

Narrow Bands

Abstract. Hyperspectral data recorded by future earth observation satellites will have up to hundreds of narrow bands that cover a wide range of the electromagnetic spectrum. The spatial resolution (around 30 meters) of such data, however, can impede the integration of the spatial domain for a classification due to spectrally mixed pixels and blurred edges in the data. Hence, the ability of performing a meaningful classification only relying on spectral information is important. In this study, a model for the spectral classification of hyperspectral data is derived by strategically optimizing a convolutional neural network (1D-CNN). The model is pre-trained and optimized on imagery of different nuts, beans, peas and dried fruits recorded with the Cubert ButterflEye X2 sensor. Subsequently, airborne hyperspectral datasets (Greding, Indian Pines and Pavia University) are used to evaluate the CNN's capability of transfer learning. For that, the datasets are classified with the pre-trained weights and, for comparison, with the same model architecture but trained from scratch with random weights. The results show substantial differences in classification accuracies (from 71.8% to 99.8% overall accuracy) throughout the used datasets, mainly caused by variations in the number of training samples, the spectral separability of the classes as well as the existence of mixed pixels for one dataset. For the dataset that is classified least accurately, the greatest improvement with pre-training is achieved (difference of 3.3% in overall accuracy compared to the non-pre-trained model). For the dataset that is classified with the highest accuracy, no significant transfer learning was observed.

Download Full-text