Model-Based Synthetic Geoelectric Sampling for Magnetotelluric Inversion With Deep Neural Networks

Deep neural networks, albeit their great success on feature learning in various computer vision tasks, are usually considered as impractical for online visual tracking, because they require very long training time and a large number of training samples. In this paper, we present an efficient and very robust tracking algorithm using a single convolutional neural network (CNN) for learning effective feature representations of the target object in a purely online manner. Our contributions are multifold. First, we introduce a novel truncated structural loss function that maintains as many training samples as possible and reduces the risk of tracking error accumulation. Second, we enhance the ordinary stochastic gradient descent approach in CNN training with a robust sample selection mechanism. The sampling mechanism randomly generates positive and negative samples from different temporal distributions, which are generated by taking the temporal relations and label noise into account. Finally, a lazy yet effective updating scheme is designed for CNN training. Equipped with this novel updating algorithm, the CNN model is robust to some long-existing difficulties in visual tracking, such as occlusion or incorrect detections, without loss of the effective adaption for significant appearance changes. In the experiment, our CNN tracker outperforms all compared state-of-the-art methods on two recently proposed benchmarks, which in total involve over 60 video sequences. The remarkable performance improvement over the existing trackers illustrates the superiority of the feature representations, which are learned purely online via the proposed deep learning framework.

Convolutional neural networks (CNNs) have been successfully applied to the single target tracking task in recent years. Generally, training a deep CNN model requires numerous labeled training samples, and the number and quality of these samples directly affect the representational capability of the trained model. However, this approach is restrictive in practice, because manually labeling such a large number of training samples is time-consuming and prohibitively expensive. In this article, we propose an active learning method for deep visual tracking, which selects and annotates the unlabeled samples to train the deep CNN model. Under the guidance of active learning, the tracker based on the trained deep CNN model can achieve competitive tracking performance while reducing the labeling cost. More specifically, to ensure the diversity of selected samples, we propose an active learning method based on multiframe collaboration to select those training samples that should be and need to be annotated. Meanwhile, considering the representativeness of these selected samples, we adopt a nearest-neighbor discrimination method based on the average nearest-neighbor distance to screen isolated samples and low-quality samples. Therefore, the training samples' subset selected based on our method requires only a given budget to maintain the diversity and representativeness of the entire sample set. Furthermore, we adopt a Tversky loss to improve the bounding box estimation of our tracker, which can ensure that the tracker achieves more accurate target states. Extensive experimental results confirm that our active-learning-based tracker (ALT) achieves competitive tracking accuracy and speed compared with state-of-the-art trackers on the seven most challenging evaluation benchmarks. Project website: https://sites.google.com/view/altrack/.

Availability of limited training remote sensing datasets is one of the problems in deep learning, as deep architectures require a large number of training samples for proper training. In this paper, we present a technique for data augmentation based on a spectral indexed generative adversarial network to train deep convolutional neural networks. This technique uses the spectral characteristic of multispectral (MS) images to support data augmentation in order to generate realistic training samples with respect to each land-use and land-cover class. The impact of multispectral remote sensing data generated through the spectral indexed GAN are evaluated through classification experiments. Experimental results obtained on the classification of the Sentinel-2 Eurosatallband datasets show that data augmentation through spectral indexed GAN enhances the main accuracy metrics.

The traditional learning based static voltage stability monitoring methods require manual labeling of a large number of training samples. Getting these labeled training sets is expensive and time-consuming in practice. To address this issue, we propose a novel disagreement-based deep learning approach for static voltage stability monitoring. The proposed method trains a deep neural network (DNN) with a small amount of labeled data to monitor the voltage stability. Case studies on the IEEE benchmark systems show that the proposed disagreement-based deep learning achieves more than 94.03% of accuracy with only 300 labeled samples. Moreover, to deal with the time variation of network operating conditions and/or network topology, we further propose a transfer learning method that can adapt the trained DNN to the new environment swiftly. The transfer process only involves a small number of labeled samples in the new environment. Our case studies show that after the network topology changes due to line tripping, the transferred DNN achieves more than 95.48% of accuracy in the new environment by utilizing only 5 labeled samples for each category.

Magnetotelluric (MT) inversion algorithms are an essential tool in exploration geophysics because they provide us with resistivity models of the subsurface. Inverse problems are ill-posed and non-unique. Resistivity models produced by geologically unconstrained MT inversions are typically smooth, which means that the geological interpretations obtained based on the resistivity models will be ambiguous and uncertain. To solve the above problems, geophysicists have made many efforts over the years to introduce any prior information (such as geological structure information and physical property data) into models, which could help to better constrain the inverted model with the inversion algorithms. In the last few years, there has been renewed interest in machine learning techniques. Various machine learning methods have been applied to inversions. Using the fuzzy c-means (FCM) clustering method, geophysicists proposed new inversion algorithms that are capable of building petrophysical information into the inversion. Although the FCM has been used and advanced in previous work, geophysicists concluded that a priori information of the correct number and value of cluster centers is very important for a successful FCM inversion. In fact, it is difficult to obtain the appropriate clustering information in some geophysical survey areas. In this study, we present an effective way to build the petrophysical information for the MT inversion based on FCM clustering algorithm. When the actual petrophysical information is insufficient, considering the characteristics of MT data, we perform a one-dimensional blocky inversion that can clearly identify the interface with distinct electrical property contrast; then, we obtain the number and value of cluster centers from 1D blocky inversion for the further MT inversion with FCM clustering. The algorithm uses guided FCM clustering to improve the model within the iterative minimization during MT inversion. In our MT inversion method, we integrate the geophysical inversion and geological differentiation into a unified scheme, which interact and enhance each other. The resistivity models obtained from the inversion respect the geophysical and petrophysical data and are easy to geologically interpret. We tested the algorithm using two synthetic examples and a field data example.

Learning-enabled spectroscopic analysis, promising for automated real-time analysis of chemicals, is facing several challenges. First, a typical machine learning model requires a large number of training samples that physical systems can not provide. Second, it requires the testing samples to be in range with the training samples, which often is not the case in the real world. Further, a spectroscopy device is limited by its memory size, computing power, and battery capacity. That requires highly efficient learning models for on-site analysis. In this paper, by analyzing multi-gas mixtures and multi-molecule suspensions, we first show that orders of magnitude reduction of data dimension can be achieved as the number of principal components that need to be retained is the same as the independent constituents in the mixture. From this principle, we designed highly compact models in which the essential principal components can be directly extracted from the interrelations between the individual chemical properties and principal components; and only a few training samples are required. Our model can predict the constituent concentrations that have not been seen in the training dataset and provide estimations of measurement noises. This approach can be extended as an effectively standardized method for principle component extraction.

Based on margin sampling (MS) strategy, an active learning approach was introduced for proposed sample selection from large quantities of labeled samples using a Landsat-7 ETM+ image to solve remote sensing image classification problems for large number of training samples. As a breakthrough from conventional random sampling and stratified systematic sampling methods, this approach ensures classification of only using a few hundred training samples to be as effective as that of using several thousand and even tens of thousands of samples by conventional methods, thereby avoiding enormous calculations, substantially reducing operating time and improving training efficiency. The test results of the proposed approach was compared with those of random sampling and stratified systematic sampling, and the effects of training samples on classification under optimized and non-optimized selection conditions was analyzed.

The field of reinforcement learning has been significantly advanced by the application of deep learning. The Deep Deterministic Policy Gradient(DDPG), an actor-critic method for continuous control, can derive satisfactory policies by use of a deep neural network. However, in common with other deep neural networks, the DDPG requires a large number of training samples and careful hyperparameter tuning.

Convolutional neural networks (CNNs) have shown results superior to most traditional image understanding approaches in many fields, incl. crop classification from satellite time series images. However, CNNs require a large number of training samples to properly train the network. The process of collecting and labeling such samples using traditional methods can be both, time-consuming and costly. To address this issue and improve classification accuracy, generating virtual training labels (VTL) from existing ones is a promising solution. To this end, this study proposes a novel method for generating VTL based on sub-dividing the training samples of each crop using self-organizing maps (SOM), and then assigning labels to a set of unlabeled pixels based on the distance to these sub-classes. We apply the new method to crop classification from Sentinel images. A three-dimensional (3D) CNN is utilized for extracting features from the fusion of optical and radar time series. The results of the evaluation show that the proposed method is effective in generating VTL, as demonstrated by the achieved overall accuracy (OA) of 95.3% and kappa coefficient (KC) of 94.5%, compared to 91.3% and 89.9% for a solution without VTL. The results suggest that the proposed method has the potential to enhance the classification accuracy of crops using VTL.

Recently, Convolutional Neural Network (CNN) has achieved some success in synthetic aperture radar (SAR) image classification. This outstanding performance mainly depends on a large number of training samples, but achieved satisfactory classification results with limited training samples remains a challenge. To address this problem, we propose a dual-scale Siamese densely connected network with Markov Random Fields (DS-SDCNet-MRF) for single-polarization SAR image classification. First, the Siamese densely connected network (SDCNet) is proposed to fully extract discriminative features under limited samples. Then, the proposed DS-SDCNet is constructed with two SDCNets of different scales to produce complementary classification results. Among them, the large-scale SDCNet has a better classification result in the homogeneous region, while the small-scale SDCNet tends to provide good detail preservation. Finally, an improved MRF model which combined the category probability information is proposed to further improve the classification performance. Experimental results on simulated and real single-polarization SAR data demonstrate that the proposed method achieves more encouraging classification performance than the current state-of-the-art classification methods with limited samples.

In High Frequency (HF) hybrid sky-surface wave radars, the spread-Doppler clutter (SDC) severely deteriorates the detection performance of low-velocity surface vessels. To suppress SDC, a large number of training samples are usually needed which can be hardly satisfied due to the heterogeneous clutter environment. In this paper, a space-frequency cascaded approach is proposed to suppress SDC using training samples from only one range cell. The proposed method involves two components. First, a frequency domain minimum variance distortionless response (MVDR) weight is designed to cancel the SDC. Second, the spatial orthogonal projection matrix is constructed to estimate clutter covariance matrix, where the signal of interest has been excluded from the training data. According to the experimental results, the proposed method can obtain a more accurate estimation of the clutter characteristics with limited training data and thus achieve better clutter suppression performance.

Erasing-inpainting-based data augmentation using denoising diffusion probabilistic models with limited samples for generalized surface defect inspection

Widely used traditional supervised deep learning methods require a large number of training samples but often fail to generalize on unseen datasets. Therefore, a more general application of any trained model is quite limited for medical imaging for clinical practice. Using separately trained models for each unique lesion category or a unique patient population will require sufficiently large curated datasets, which is not practical to use in a real-world clinical set-up. Few-shot learning approaches can not only minimize the need for an enormous number of reliable ground truth labels that are labour-intensive and expensive, but can also be used to model on a dataset coming from a new population. To this end, we propose to exploit an optimization-based implicit model agnostic meta-learning (iMAML) algorithm under few-shot settings for medical image segmentation. Our approach can leverage the learned weights from diverse but small training samples to perform analysis on unseen datasets with high accuracy. We show that, unlike classical few-shot learning approaches, our method improves generalization capability. To our knowledge, this is the first work that exploits iMAML for medical image segmentation and explores the strength of the model on scenarios such as meta-training on unique and mixed instances of lesion datasets. Our quantitative results on publicly available skin and polyp datasets show that the proposed method outperforms the naive supervised baseline model and two recent few-shot segmentation approaches by large margins. In addition, our iMAML approach shows an improvement of 2%–4% in dice score compared to its counterpart MAML for most experiments.

The effective extraction of impervious surfaces is critical to monitor their expansion and ensure the sustainable development of cities. Open geographic data can provide a large number of training samples for machine learning methods based on remote-sensed images to extract impervious surfaces due to their advantages of low acquisition cost and large coverage. However, training samples generated from open geographic data suffer from severe sample imbalance. Although one-class methods can effectively extract an impervious surface based on imbalanced samples, most of the current one-class methods ignore the fact that an impervious surface comprises varied geographic objects, such as roads and buildings. Therefore, this paper proposes an object-oriented deep multi-sphere support vector data description (OODMSVDD) method, which takes into account the diversity of impervious surfaces and incorporates a variety of open geographic data involving OpenStreetMap (OSM), Points of Interest (POIs), and trajectory GPS points to automatically generate massive samples for model learning, thereby improving the extraction of impervious surfaces with varied types. The feasibility of the proposed method is experimentally verified with an overall accuracy of 87.43%, and its superior impervious surface classification performance is shown via comparative experiments. This provides a new, accurate, and more suitable extraction method for complex impervious surfaces.

Active Shape Models (ASM) have been applied to various segmentation tasks in medical imaging, most successfully in 2D segmentation of objects that have a fairly consistent shape. However, several difficulties arise when extending 2D ASM to 3D: (1) difficulty in 3D labeling, (2) the requirement of a large number of training samples, (3) the challenging problem of landmark correspondence in 3D, (4) inefficient initialization and optimization in 3D. This paper addresses the 3D segmentation problem by using a small number of effective 2D statistical models called oriented ASM (OASM). We demonstrate that a small number of 2D OASM models, which are derived from a chunk of a contiguous set of slices, are sufficient to capture the shape variation between slices and individual objects. Each model can be matched rapidly to a new slice by using the OASM algorithm¹. Our experiments in segmenting breast and bone of the foot in MR images indicate the following: (1) The accuracy of segmentation via our method is much better than that of 2DASM-based segmentation methods.² (2) Far fewer landmarks are required compared with thousands of landmarks needed in true 3D ASM. Therefore, far fewer training samples are required to capture details. (3) Our method is computationally slightly more expensive than the 2D method2 owing to its 2 level dynamic programming (2LDP) algorithm.