ABSTRACT
Bone proliferation is an important pathological feature of inflammatory rheumatic diseases. Although recent advance in high-resolution peripheral quantitative computed tomography (HR-pQCT) enables physicians to study microarchitectures, physicians' annotation of proliferation suffers from slice inconsistency and subjective variations. Also, there are only few effective automatic or semi-automatic tools for proliferation detection. In this study, by integrating pathological knowledge of proliferation formation with the advancement of statistical shape analysis theory, we present an unsupervised method, named Deformation-Controllable Elastic Shape model, for 3D bone Proliferation Analysis (DCES-PA). Unlike previous shape analysis methods that directly regularize the smoothness of the displacement field, DCES-PA regularizes the first and second-order derivative of the displacement field and decomposes these vector fields according to different deformations. For the first-order elastic metric, DCES-PA orthogonally decomposes the first-order derivative of the displacement field by shearing, scaling and bending deformation, and then penalize deformations triggering proliferation formation. For the second-order elastic metric, DCES-PA encodes both intrinsic and extrinsic surface curvatures into the second-order derivative of the displacement field to control the generation of high-curvature regions. By integrating the elastic shape metric with the varifold distances, DCES-PA achieves correspondence-free shape analysis. Extensive experiments on both simulated and real clinical datasets demonstrate that DCES-PA not only shows an improved accuracy than other state-of-the-art shape-based methods applied to proliferation analysis but also produces highly sensitive proliferation annotations to assist physicians in proliferation analysis.
Subject(s)
Imaging, Three-Dimensional , Tomography, X-Ray Computed , Humans , Tomography, X-Ray Computed/methods , Imaging, Three-Dimensional/methods , Bone and Bones/diagnostic imaging , Hand/diagnostic imaging , Female , Male , Cell ProliferationABSTRACT
Human emotions contain both basic and compound facial expressions. In many practical scenarios, it is difficult to access all the compound expression categories at one time. In this paper, we investigate comprehensive facial expression recognition (FER) in the class-incremental learning paradigm, where we define well-studied and easily-accessible basic expressions as initial classes and learn new compound expressions incrementally. To alleviate the stability-plasticity dilemma in our incremental task, we propose a novel Relationship-Guided Knowledge Transfer (RGKT) method for class-incremental FER. Specifically, we develop a multi-region feature learning (MFL) module to extract fine-grained features for capturing subtle differences in expressions. Based on the MFL module, we further design a basic expression-oriented knowledge transfer (BET) module and a compound expression-oriented knowledge transfer (CET) module, by effectively exploiting the relationship across expressions. The BET module initializes the new compound expression classifiers based on expression relevance between basic and compound expressions, improving the plasticity of our model to learn new classes. The CET module transfers expression-generic knowledge learned from new compound expressions to enrich the feature set of old expressions, facilitating the stability of our model against forgetting old classes. Extensive experiments on three facial expression databases show that our method achieves superior performance in comparison with several state-of-the-art methods.
Subject(s)
Facial Recognition , Humans , Emotions , Learning , Facial Expression , Databases, FactualABSTRACT
Graph convolutional networks (GCN) have recently been studied to exploit the graph topology of the human body for skeleton-based action recognition. However, most of these methods unfortunately aggregate messages via an inflexible pattern for various action samples, lacking the awareness of intra-class variety and the suitableness for skeleton sequences, which often contain redundant or even detrimental connections. In this paper, we propose a novel Deformable Graph Convolutional Network (DeGCN) to adaptively capture the most informative joints. The proposed DeGCN learns the deformable sampling locations on both spatial and temporal graphs, enabling the model to perceive discriminative receptive fields. Notably, considering human action is inherently continuous, the corresponding temporal features are defined in a continuous latent space. Furthermore, we design an innovative multi-branch framework, which not only strikes a better trade-off between accuracy and model size, but also elevates the effect of ensemble between the joint and bone modalities remarkably. Extensive experiments show that our proposed method achieves state-of-the-art performances on three widely used datasets, NTU RGB+D, NTU RGB+D 120, and NW-UCLA.
ABSTRACT
Knowledge distillation (KD), which aims at transferring the knowledge from a complex network (a teacher) to a simpler and smaller network (a student), has received considerable attention in recent years. Typically, most existing KD methods work on well-labeled data. Unfortunately, real-world data often inevitably involve noisy labels, thus leading to performance deterioration of these methods. In this article, we study a little-explored but important issue, i.e., KD with noisy labels. To this end, we propose a novel KD method, called ambiguity-guided mutual label refinery KD (AML-KD), to train the student model in the presence of noisy labels. Specifically, based on the pretrained teacher model, a two-stage label refinery framework is innovatively introduced to refine labels gradually. In the first stage, we perform label propagation (LP) with small-loss selection guided by the teacher model, improving the learning capability of the student model. In the second stage, we perform mutual LP between the teacher and student models in a mutual-benefit way. During the label refinery, an ambiguity-aware weight estimation (AWE) module is developed to address the problem of ambiguous samples, avoiding overfitting these samples. One distinct advantage of AML-KD is that it is capable of learning a high-accuracy and low-cost student model with label noise. The experimental results on synthetic and real-world noisy datasets show the effectiveness of our AML-KD against state-of-the-art KD methods and label noise learning (LNL) methods. Code is available at https://github.com/Runqing-forMost/ AML-KD.
ABSTRACT
GAN inversion aims to invert a given image back into the latent space of a pretrained GAN model so that the image can be faithfully reconstructed from the inverted code by the generator. As an emerging technique to bridge the real and fake image domains, GAN inversion plays an essential role in enabling pretrained GAN models, such as StyleGAN and BigGAN, for applications of real image editing. Moreover, GAN inversion interprets GAN's latent space and examines how realistic images can be generated. In this paper, we provide a survey of GAN inversion with a focus on its representative algorithms and its applications in image restoration and image manipulation. We further discuss the trends and challenges for future research. A curated list of GAN inversion methods, datasets, and other related information can be found at https://github.com/weihaox/awesome-gan-inversion.
ABSTRACT
Rheumatoid arthritis (RA) is a chronic inflammatory disease. It leads to bone erosion in joints and other complications, which severely affect patients' quality of life. To accurately diagnose and monitor the progression of RA, quantitative imaging and analysis tools are desirable. High-resolution peripheral quantitative computed tomography (HR-pQCT) is such a promising tool for monitoring disease progression in RA. However, automatic erosion detection tools using HR-pQCT images are not yet available. Inspired by the consensus among radiologists on the erosions in HR-pQCT images, in this paper we define erosion as the significant concave regions on the cortical layer, and develop a model-based 3D automatic erosion detection method. It mainly consists of two steps: constructing closed cortical surface, and detecting erosion regions on the surface. In the first step, we propose an initialization-robust region competition methods for joint segmentation, and then fill the surface gaps by using joint bone separation and curvature-based surface alignment. In the second step, we analyze the curvature information of each voxel, and then aggregate the candidate voxels into concave surface regions and use the shape information of the regions to detect the erosions. We perform qualitative assessments of the new method using 59 well-annotated joint volumes. Our method has shown satisfactory and consistent performance compared with the annotations provided by medical experts.
Subject(s)
Arthritis, Rheumatoid , Quality of Life , Humans , Tomography, X-Ray Computed/methods , Arthritis, Rheumatoid/diagnostic imaging , HandABSTRACT
Person attribute recognition (PAR) aims to simultaneously predict multiple attributes of a person. Existing deep learning-based PAR methods have achieved impressive performance. Unfortunately, these methods usually ignore the fact that different attributes have an imbalance in the number of noisy-labeled samples in the PAR training datasets, thus leading to suboptimal performance. To address the above problem of imbalanced noisy-labeled samples, we propose a novel and effective loss called drop loss for PAR. In the drop loss, the attributes are treated differently in an easy-to-hard way. In particular, the noisy-labeled candidates, which are identified according to their gradient norms, are dropped with a higher drop rate for the harder attribute. Such a manner adaptively alleviates the adverse effect of imbalanced noisy-labeled samples on model learning. To illustrate the effectiveness of the proposed loss, we train a simple ResNet-50 model based on the drop loss and term it DropNet. Experimental results on two representative PAR tasks (including facial attribute recognition and pedestrian attribute recognition) demonstrate that the proposed DropNet achieves comparable or better performance in terms of both balanced accuracy and classification accuracy over several state-of-the-art PAR methods.
ABSTRACT
In this paper, we reveal the discriminant capacity of orthogonal data projection onto the generalized difference subspace (GDS), both theoretically and experimentally. In our previous work, we demonstrated that the GDS projection works as a quasi-orthogonalization of class subspaces, which is an effective feature extraction for subspace based classifiers. Here, we further show that GDS projection also works as a discriminant feature extraction through a similar mechanism to the Fisher discriminant analysis (FDA). A direct proof of the connection between GDS projection and FDA is difficult due to the significant difference in their formulations. To circumvent the complication, we first introduce geometrical Fisher discriminant analysis (gFDA) based on a simplified Fisher criterion. It is derived from a heuristic yet practically plausible assumption: the direction of the sample mean vector of a class is largely aligned to the first principal component vector of the class, given that the principal component analysis (PCA) is applied without data centering. gFDA works stably even under few samples, bypassing the small sample size (SSS) problem of FDA. We then prove that gFDA is equivalent to GDS projection with a small correction term. This equivalence ensures GDS projection to inherit the discriminant ability from FDA via gFDA. Furthermore, we discuss two useful extensions of these methods, 1) a nonlinear extension by kernel trick, 2) a combination with CNN features. The equivalence and the effectiveness of the extensions have been verified through extensive experiments on the extended Yale B+, CMU face database, ALOI, ETH80, MNIST, and CIFAR10, mainly focusing on image recognition under small samples.
ABSTRACT
As a typical non-Gaussian vector variable, a neutral vector variable contains nonnegative elements only, and its l1 -norm equals one. In addition, its neutral properties make it significantly different from the commonly studied vector variables (e.g., the Gaussian vector variables). Due to the aforementioned properties, the conventionally applied linear transformation approaches [e.g., principal component analysis (PCA) and independent component analysis (ICA)] are not suitable for neutral vector variables, as PCA cannot transform a neutral vector variable, which is highly negatively correlated, into a set of mutually independent scalar variables and ICA cannot preserve the bounded property after transformation. In recent work, we proposed an efficient nonlinear transformation approach, i.e., the parallel nonlinear transformation (PNT), for decorrelating neutral vector variables. In this article, we extensively compare PNT with PCA and ICA through both theoretical analysis and experimental evaluations. The results of our investigations demonstrate the superiority of PNT for decorrelating the neutral vector variables.
ABSTRACT
Noise annoyance has been often reported as one of the main adverse effects of noise exposure on human health, and there is consensus that it relates to several factors going beyond the mere energy content of the signal. Research has historically focused on a limited set of sound sources (e.g., transport and industrial noise); only more recently is attention being given to more holistic aspects of urban acoustic environments and the role they play in the noise annoyance perceptual construct. This is the main approach promoted in soundscape studies, looking at both wanted and unwanted sounds. In this study, three specific aspects were investigated, namely: (1) the effect of different sound sources combinations, (2) the number of sound sources present in the soundscape, and (3) the presence of individual sound source, on noise annoyance perception. For this purpose, a large-scale online experiment was carried out with 1.2k+ participants, using 2.8k+ audio recordings of complex urban acoustic environments to investigate how they would influence the perceived noise annoyance. Results showed that: (1) the combinations of different sound sources were not important, compared, instead, to the number of sound sources identified in the soundscape recording (regardless of sound sources type); (2) the annoyance ratings expressed a minimum when any two clearly distinguishable sound sources were present in a given urban soundscape; and (3) the presence (either in isolation or combination) of traffic-related sound sources increases noise annoyance, while the presence (either in isolation or combination) of nature-related sound sources decreases noise annoyance.
Subject(s)
Auditory Perception , Noise , Humans , Noise/adverse effects , Sound , Acoustics , IndustryABSTRACT
Metric learning aims to learn a distance metric such that semantically similar instances are pulled together while dissimilar instances are pushed away. Many existing methods consider maximizing or at least constraining a distance margin in the feature space that separates similar and dissimilar pairs of instances to guarantee their generalization ability. In this article, we advocate imposing an adversarial margin in the input space so as to improve the generalization and robustness of metric learning algorithms. We first show that the adversarial margin, defined as the distance between training instances and their closest adversarial examples in the input space, takes account of both the distance margin in the feature space and the correlation between the metric and triplet constraints. Next, to enhance robustness to instance perturbation, we propose to enlarge the adversarial margin through minimizing a derived novel loss function termed the perturbation loss. The proposed loss can be viewed as a data-dependent regularizer and easily plugged into any existing metric learning methods. Finally, we show that the enlarged margin is beneficial to the generalization ability by using the theoretical technique of algorithmic robustness. Experimental results on 16 datasets demonstrate the superiority of the proposed method over existing state-of-the-art methods in both discrimination accuracy and robustness against possible noise.
ABSTRACT
Recent methods in network pruning have indicated that a dense neural network involves a sparse subnetwork (called a winning ticket), which can achieve similar test accuracy to its dense counterpart with much fewer network parameters. Generally, these methods search for the winning tickets on well-labeled data. Unfortunately, in many real-world applications, the training data are unavoidably contaminated with noisy labels, thereby leading to performance deterioration of these methods. To address the above-mentioned problem, we propose a novel two-stream sample selection network (TS 3 -Net), which consists of a sparse subnetwork and a dense subnetwork, to effectively identify the winning ticket with noisy labels. The training of TS 3 -Net contains an iterative procedure that switches between training both subnetworks and pruning the smallest magnitude weights of the sparse subnetwork. In particular, we develop a multistage learning framework including a warm-up stage, a semisupervised alternate learning stage, and a label refinement stage, to progressively train the two subnetworks. In this way, the classification capability of the sparse subnetwork can be gradually improved at a high sparsity level. Extensive experimental results on both synthetic and real-world noisy datasets (including MNIST, CIFAR-10, CIFAR-100, ANIMAL-10N, Clothing1M, and WebVision) demonstrate that our proposed method achieves state-of-the-art performance with very small memory consumption for label noise learning. Code is available at https://github.com/Runqing-forMost/TS3-Net/tree/master.
ABSTRACT
This article proposes a survival model based on graph convolutional networks (GCNs) with geometric graphs directly constructed from high-dimensional features. First, we clarify that the graphs used in GCNs play an important role in processing the relational information of samples, and the graphs that align well with the underlying data structure could be beneficial for survival analysis. Second, we show that sparse geometric graphs derived from high-dimensional data are more favorable compared with dense graphs when used in GCNs for survival analysis. Third, from this insight, we propose a model for survival analysis based on GCNs. By using multiple sparse geometric graphs and a proposed sequential forward floating selection algorithm, the new model is able to simultaneously perform survival analysis and unveil the local neighborhoods of samples. The experimental results on real-world datasets show that the proposed survival analysis approach based on GCNs outperforms a variety of existing methods and indicate that geometric graphs can aid survival analysis of high-dimensional data.
ABSTRACT
We propose a polarization-based probabilistic discriminative model for deriving a set of new sigmoid-transformed polarimetry feature parameters, which not only enables accurate and quantitative characterization of cancer cells at pixel level, but also accomplish the task with a simple and stable model. By taking advantages of polarization imaging techniques, these parameters enable a low-magnification and wide-field imaging system to separate the types of cells into more specific categories that previously were distinctive under high magnification. Instead of blindly choosing the model, the L0 regularization method is used to obtain the simplified and stable polarimetry feature parameter. We demonstrate the model viability by using the pathological tissues of breast cancer and liver cancer, in each of which there are two derived parameters that can characterize the cells and cancer cells respectively with satisfactory accuracy and sensitivity. The stability of the final model opens the possibility for physical interpretation and analysis. This technique may bypass the typically labor-intensive and subjective tumor evaluating system, and could be used as a blueprint for an objective and automated procedure for cancer cell screening.
ABSTRACT
BACKGROUND: The timely, rapid, and accurate near real-time observations are urgent to monitor the damage of corn armyworm, because the rapid expansion of armyworm would lead to severe yield losses. Therefore, the potential of machine learning algorithms for identifying the armyworm infected areas automatically and accurately by multispectral unmanned aerial vehicle (UAV) dataset is explored in this study. The study area is in Beicuizhuang Village, Langfang City, Hebei Province, which is the main corn-producing area in the North China Plain. RESULTS: Firstly, we identified the optimal combination of image features by Gini-importance and the comparation of four kinds of machine learning methods including Random Forest (RF), Multilayer Perceptron (MLP), Naive Bayesian (NB) and Support Vector Machine (SVM) was done. And RF was proved to be the most potential with the highest Kappa and OA of 0.9709 and 0.9850, respectively. Secondly, the armyworm infected areas and healthy corn areas were predicted by an optimized RF model in the UAV dataset, and the armyworm incidence levels were classified subsequently. Thirdly, the relationship between the spectral characteristics of different bands and pest incidence levels within the Sentinel-2 and UAV images were analyzed, and the B3 in UAV images and the B6 in Sentinel-2 image were less sensitive for armyworm incidence levels. Therefore, the Sentinel-2 image was used to monitor armyworm in two towns. CONCLUSIONS: The optimized dataset and RF model are effective and reliable, which can be used for identifying the corn damage by armyworm using UAV images accurately and automatically in field-scale. © 2022 Society of Chemical Industry.
Subject(s)
Unmanned Aerial Devices , Zea mays , Animals , Bayes Theorem , Seasons , SpodopteraABSTRACT
Different from the object motion blur, the defocus blur is caused by the limitation of the cameras' depth of field. The defocus amount can be characterized by the parameter of point spread function and thus forms a defocus map. In this paper, we propose a new network architecture called Defocus Image Deblurring Auxiliary Learning Net (DID-ANet), which is specifically designed for single image defocus deblurring by using defocus map estimation as auxiliary task to improve the deblurring result. To facilitate the training of the network, we build a novel and large-scale dataset for single image defocus deblurring, which contains the defocus images, the defocus maps and the all-sharp images. To the best of our knowledge, the new dataset is the first large-scale defocus deblurring dataset for training deep networks. Moreover, the experimental results demonstrate that the proposed DID-ANet outperforms the state-of-the-art methods for both tasks of defocus image deblurring and defocus map estimation, both quantitatively and qualitatively. The dataset, code, and model is available on GitHub: https://github.com/xytmhy/DID-ANet-Defocus-Deblurring.
ABSTRACT
The success of deep convolutional networks (ConvNets) generally relies on a massive amount of well-labeled data, which is labor-intensive and time-consuming to collect and annotate in many scenarios. To eliminate such limitation, self-supervised learning (SSL) is recently proposed. Specifically, by solving a pre-designed proxy task, SSL is capable of capturing general-purpose features without requiring human supervision. Existing efforts focus obsessively on designing a particular proxy task but ignore the semanticity of samples that are advantageous to downstream tasks, resulting in the inherent limitation that the learned features are specific to the proxy task, namely the proxy task-specificity of features. In this work, to improve the generalizability of features learned by existing SSL methods, we present a novel self-supervised framework SSL++ to incorporate the proxy task-independent semanticity of samples into the representation learning process. Technically, SSL++ aims to leverage the complementarity, between the low-level generic features learned by a proxy task and the high-level semantic features newly learned by the generated semantic pseudo-labels, to mitigate the task-specificity and improve the generalizability of features. Extensive experiments show that SSL++ performs favorably against the state-of-the-art approaches on the established and latest SSL benchmarks.
Subject(s)
Supervised Machine Learning , HumansABSTRACT
Object detection has made enormous progress and has been widely used in many applications. However, it performs poorly when only limited training data is available for novel classes that the model has never seen before. Most existing approaches solve few-shot detection tasks implicitly without directly modeling the detectors for novel classes. In this article, we propose GenDet, a new meta-learning-based framework that can effectively generate object detectors for novel classes from few shots and, thus, conducts few-shot detection tasks explicitly. The detector generator is trained by numerous few-shot detection tasks sampled from base classes each with sufficient samples, and thus, it is expected to generalize well on novel classes. An adaptive pooling module is further introduced to suppress distracting samples and aggregate the detectors generated from multiple shots. Moreover, we propose to train a reference detector for each base class in the conventional way, with which to guide the training of the detector generator. The reference detectors and the detector generator can be trained simultaneously. Finally, the generated detectors of different classes are encouraged to be orthogonal to each other for better generalization. The proposed approach is extensively evaluated on the ImageNet, VOC, and COCO data sets under various few-shot detection settings, and it achieves new state-of-the-art results.
ABSTRACT
Due to lack of data, overfitting ubiquitously exists in real-world applications of deep neural networks (DNNs). We propose advanced dropout, a model-free methodology, to mitigate overfitting and improve the performance of DNNs. The advanced dropout technique applies a model-free and easily implemented distribution with parametric prior, and adaptively adjusts dropout rate. Specifically, the distribution parameters are optimized by stochastic gradient variational Bayes in order to carry out an end-to-end training. We evaluate the effectiveness of the advanced dropout against nine dropout techniques on seven computer vision datasets (five small-scale datasets and two large-scale datasets) with various base models. The advanced dropout outperforms all the referred techniques on all the datasets. We further compare the effectiveness ratios and find that advanced dropout achieves the highest one on most cases. Next, we conduct a set of analysis of dropout rate characteristics, including convergence of the adaptive dropout rate, the learned distributions of dropout masks, and a comparison with dropout rate generation without an explicit distribution. In addition, the ability of overfitting prevention is evaluated and confirmed. Finally, we extend the application of the advanced dropout to uncertainty inference, network pruning, text classification, and regression. The proposed advanced dropout is also superior to the corresponding referred methods. Codes are available at https://github.com/PRIS-CV/AdvancedDropout.
Subject(s)
Algorithms , Neural Networks, Computer , Bayes TheoremABSTRACT
This paper proposes a dual-supervised uncertainty inference (DS-UI) framework for improving Bayesian estimation-based UI in DNN-based image recognition. In the DS-UI, we combine the classifier of a DNN, i.e., the last fully-connected (FC) layer, with a mixture of Gaussian mixture models (MoGMM) to obtain an MoGMM-FC layer. Unlike existing UI methods for DNNs, which only calculate the means or modes of the DNN outputs' distributions, the proposed MoGMM-FC layer acts as a probabilistic interpreter for the features that are inputs of the classifier to directly calculate the probabilities of them for the DS-UI. In addition, we propose a dual-supervised stochastic gradient-based variational Bayes (DS-SGVB) algorithm for the MoGMM-FC layer optimization. Unlike conventional SGVB and optimization algorithms in other UI methods, the DS-SGVB not only models the samples in the specific class for each Gaussian mixture model (GMM) in the MoGMM, but also considers the negative samples from other classes for the GMM to reduce the intra-class distances and enlarge the inter-class margins simultaneously for enhancing the learning ability of the MoGMM-FC layer in the DS-UI. Experimental results show the DS-UI outperforms the state-of-the-art UI methods in misclassification detection. We further evaluate the DS-UI in open-set out-of-domain/-distribution detection and find statistically significant improvements. Visualizations of the feature spaces demonstrate the superiority of the DS-UI. Codes are available at https://github.com/PRIS-CV/DS-UI.