Target tracking in dynamic environments faces challenges such as occlusions and appearance changes, especially long-time tracking. To address this, we propose GANMem-Siamese, integrating Generative Adversarial Networks (GANs) with Siamese networks, enhanced by a cycle memory mechanism and channel attention. The cycle memory mechanism selects key historical frames based on confidence scores, ensuring effective long-term representation. Within the selected frames, we introduce a channel attention-based feature fusion strategy, which adaptively integrates GAN-generated features with intermediate template features, enhancing feature diversity while maintaining temporal coherence. This approach mitigates feature drift and improves tracking robustness. Experimental results on benchmark datasets, including OTB100, VOT2016, and UAV123-20L, demonstrate that our method achieves state-of-the-art performance among unsupervised trackers, significantly improving long-term tracking stability and accuracy.

In order to improve the performance of Kabaddi athlete posture target tracking, a method of Kabaddi athlete posture target tracking based on machine learning is proposed. The threshold change parameter is calculated by using the obtained athletes’ posture characteristic parameters, and the golden section is introduced to transform it and smooth the features of athletes’ posture are extracted. Constraint loss is added to the local global supervision module of machine learning, and the local features of athlete pose are integrated, and the local parameters of athlete pose are obtained by loss function. Taylor formula was used to calculate the athletes’ pose velocity, and Kalman filter was used to evaluate the joint motion data, and Kabaddi athlete pose model was constructed. The frame difference of the background image is calculated by normalizing the athletes’ pose image, and the athletes’ pose is automatically tracked. The experimental results show that this method can track all nodes of the athletes’ posture, and has good performance in the absolute error, detail loss and tracking lag rate of the athletes’ posture tracking, so it is helpful to improve the athletes’ technical level and improve the training effect.

The vulnerability of Deep Neural Networks (DNNs) to adversarial perturbations has been demonstrated in a large body of research. Compared to image-dependent adversarial perturbations, universal adversarial perturbations(UAPs) is more challenging for indiscriminately attacking the model inputs. However, there are few studies on generating data-free targeted UAPs and the targeted attack success rate of the latest method remains unsatisfactory. Not only that, fewer studies have implemented their approach on Transformers and its efficacy remains uncertain. Therefore, a novel method denoted as Denoising Targeted UAP (DT-UAP) is proposed in this paper that considers the training input as the noise, and takes the input of the last layer into calculation. Specifically, the proposed method minimizes the distance between perturbations and adversarial examples, then incorporates a targeted loss function to generate targeted universal adversarial perturbations for different DNNs and Transformers based on different proxy datasets. DT-UAP has achieved an average improvement of 5% to 10% in terms of both fooling rate and targeted fooling rate comparing to the most recent method for generating targeted universal adversarial perturbation with proxy dataset for DNNs. Additionally, DT-UAP has also achieved a targeted attack success rate of over 80% on Transformers such as MaxVit and SwinTransformer.

Incremental learning for image classification has been widely studied in the past few years, but few works explored incremental learning for object detection. Most existing incremental object detectors deploy knowledge distillation to constrain the model to retain old knowledge and avoid the catastrophic forgetting phenomenon. Nonetheless, this common practice results in strong constraints adhering to the old knowledge, therefore deteriorating the learning ability to new knowledge. In this work, we propose a new framework named Contrastive R-CNN for incremental learning of object detection to balance the retaining of the old knowledge and the learning of the new knowledge. The proposed framework is mainly composed of two modules, data distillation and temporal contrast. Data distillation presents a median entropy filter strategy to generate the annotations for the RoIs of the old objects, while temporal contrast designs an RoI contrast mechanism to minimize the ambiguity between old and new instances for better incremental learning. Extensive experiments on the PASCAL VOC dataset demonstrate the effectiveness of our proposed approach.

Unsupervised clustering categorizes a sample set into several groups, where the samples in the same group share high-level concepts. As the clustering performances are heavily determined by the metric to assess the similarity between sample pairs, we propose to learn a deep similarity score function and use it to capture the correlations between sample pairs for improved clustering. We formulate the learn-ing procedure in a ranking framework and introduce two new supervisory signals to train our model. Specifically, we train the similarity score function to guarantee 1) a sample should have a higher level of similarity with its nearest neighbors than others in order to achieve correct clustering, and 2) the ordering of the similarity between neighboring sample pairs should be preserved in order to achieve robust clustering. To this end, we not only study the relevance between neighboring sample pairs for lo -cal structure learning, but also study the relevance between each sample and the boundary samples for global structure learning. Extensive experiments on seven public available datasets validate the effective-ness of our proposed framework, including face image clustering, object image clustering, and real-world image clustering.(c) 2022 Elsevier Ltd. All rights reserved.

Deep learning has achieved remarkable progress in computer vision and image analysis. However, raindrop removal from single image still remains challenging, due to a wide range of raindrop diversities and surface reflections. In this paper, we propose an iterative neural network with feedback strategy and contrastive learning for single image raindrop removal. First, we design an iterative feedback neural network to refine low-level representations with high-level information, i.e., the output of the previous iteration is used as input for the next iteration, together with the input image with raindrops. As a result, raindrops could be gradually removed through this feedback manner. Then, we deploy contrastive regularization to push the restored image from each iteration close to the clean images without raindrops, but away from rainy images with raindrops. Extensive experiments on two raindrop benchmark datasets demonstrate the effectiveness of the proposed approach in comparison with the state-of-the-art methods. The methodology in this work could be further extended to self-supervised contrastive learning to obtain robust feature representations with less labelled data.

In the existing research on Chinese natural language processing text feature extraction has always been a core problem in this field, and the advantages and disadvantages of text feature extraction performance are important factors that directly affect the performance of text processing technology. This research offers a dual-channel feature extraction method based on neural networks in order to handle this challenge of text feature extraction from the perspectives of local information and context global information of text sentences. The text's local features are extracted using GCN, while the text's overall semantic features are extracted using BiLSTM, and the full connected neural network is used to fuse the local eigenvector with the global eigenvector to obtain the final features of the text. To confirm the efficiency of the suggested feature extraction technique, the suggested approach is used to classify news material in this research. The outcomes of the experiments demonstrate that the performance of the suggested method outperforms that of the GCN and BiLSTM feature extraction methods used independently. When compared to other feature extraction approaches, this method's experimental results are still better.

Conventional discriminative-correlation-filter-based (DCF-based) visual tracking methods always update model at a fixed frequency and learning rate. Without evaluating the tracking confidence scores, the response map generated by filter is the only evidence for locating. Thus, most of the existing DCF-based methods suffer from the model contamination caused by drastic appearance variations, which leads to tracking drift even failure. And excessively frequent update will increase the computational redundancy and risk of over-fitting. In addition, these methods cannot recover target from heavy occlusion neither. Based on the observation that the shape of response maps reflects the matching degree between filter and target, we design and train a small-scale binary network named as response map analysis network (RAN) to evaluate the confidence scores of filters. Further, we propose to learn multiple filters to exploit different kinds of features, and adaptively adjust the update parameters according to the corresponding confidence scores. Moreover, we build a simple occlusion event model to detect heavy occlusion and recover target. Extensive experimental results validate the effectiveness of RAN and demonstrate that the proposed tracker performs favorably against other state-of-the-art (SOTA) DCF-based trackers in terms of precision, overlap rate and efficiency. • We propose an effective way of evaluating the tracking confidence score of correlation filter utilizing CNN. • We propose to learn multiple filters to exploit different features and adaptively adjust the update parameters according to confidence scores. • We build a simple occlusion event model to detect heavy occlusion and recover target. • The improved DCF tracker performs favorably compared to other SOTA trackers on several large-scale benchmarks.

Considering the problems of similarity interference, partial occlusions, and changes in scale during target tracking, a target tracking method based on interference detection is proposed, which is an improvement over the Siamese fully convolutional classification and regression neural network (SiamCAR) approach. Under the proposed framework, the marginal distribution of the feature maps is used to determine the presence or absence of interferents. When interference is present in a scene, a motion vector composed of the predicted value obtained through a Kalman filter is used as the basis for target prediction. Experiments on the benchmark LaSOT dataset show that the proposed algorithm based on SiamCAR, which introduces motion features, achieves the best performance in videos with similar object interference, partial occlusions, fast motion, and small target tracking, as compared with the classical SiamCAR and other excellent target tracking algorithms.

Object tracking is a challenging task in computer vision. The correlation filter based trackers are widely used for visual tracking due to their efficiencies. However, they cannot handle occlusion very well. In this paper, an effective method is proposed for occlusion detection based on high-level classification scores from the Convolutional Neural Network (CNN) trained on the ImageNet dataset. Also, we propose a novel tracking method by holistically considering multiple tracking models trained previously. In each frame, multiple correlation filters are first trained using hierarchical convolutional features, and then progressively selected according to the so-called tracking quality (status). Finally, a linear motion model is adopted to effectively re-detect the lost target. Experimental results have demonstrated that our method achieved good performance for handling occlusion.

Neuroimaging studies have revealed that the structural changes of the corpus callosum (CC) are evident in a variety of neurological diseases, such as epilepsy and autism. Segmentation of the CC from magnetic resonance images (MRI) of the brain is a crucial step in the diagnosis of various brain disorders. However, the lack of open benchmark CC datasets has hindered development of CC segmentation techniques. In this work, we present an open benchmark dataset - OpenCC - for CC segmentation and evaluation. The dataset was built through alternative application of automatic segmentation and manual refinement. The automatic segmentation is based on recent advances in deep learning - fully convolutional networks, specifically U-Net, while the manual refinement is done by domain radiologists. The resulting dataset consists of 4643 mid-sagittal (or near mid-sagittal) slices and their corresponding CC masks. Furthermore, we provided some baseline segmentation results on the OpenCC dataset by using two latest deep learning segmentation approaches. The OpenCC dataset can be used for comparison and evaluation of newly developed CC segmentation algorithms. We endeavor that, through the publishing of the OpenCC dataset and baseline segmentation results, we could promote further development of CC segmentation techniques.

Credit card has become popular mode of payment for both online and offline purchase, which leads to increasing daily fraud transactions. An Efficient fraud detection methodology is therefore essential to maintain the reliability of the payment system. In this study, we perform a comparison study of credit card fraud detection by using various supervised and unsupervised approaches. Specifically, 6 supervised classification models, i.e., Logistic Regression (LR), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Tree (DT), Random Forest (RF), Extreme Gradient Boosting (XGB), as well as 4 unsupervised anomaly detection models, i.e., One-Class SVM (OCSVM), Auto-Encoder (AE), Restricted Boltzmann Machine (RBM), and Generative Adversarial Networks (GAN), are explored in this study. We train all these models on a public credit card transaction dataset from Kaggle website, which contains 492 frauds out of 284,807 transactions. The labels of the transactions are used for supervised learning models only. The performance of each model is evaluated through 5-fold cross validation in terms of Area Under the Receiver Operating Curves (AUROC). Within supervised approaches, XGB and RF obtain the best performance with AUROC = 0.989 and AUROC = 0.988, respectively. While for unsupervised approaches, RBM achieves the best performance with AUROC = 0.961, followed by GAN with AUROC = 0.954. The experimental results show that supervised models perform slightly better than unsupervised models in this study. Anyway, unsupervised approaches are still promising for credit card fraud transaction detection due to the insufficient annotation and the data imbalance issue in real-world applications.

Recently, Siamese networks have drawn great attention in visual tracking community because of their balanced accuracy and speed. By comparing the target patch with the candidate windows in a search region, we can track the object to the location where the highest similarity score is obtained. However, in Siamese Network, pairs of training data come from different frames of the same video, and for each search area, the non-semantic background occupies the majority, while semantic entities and distractor occupy less. This imbalanced distribution makes the training model hard to learn instance-level representation, but tending to learn the differences between foreground and background. For the targets those with large differences in the background also achieve high scores, and even some extraneous objects get high scores. To overcome this limitation, we enhance Siamese Network by color histogram based Bayes classifier. This method allows us to identify potentially distracting regions in advance. The risk of drifting is significantly reduced. Experiment results show that our tracker achieves state-of-the-art performance.

Recently, deep learning has achieved very promising results in visual object tracking. Deep neural networks in existing tracking methods require a lot of training data to learn a large number of parameters. However, training data is not sufficient for visual object tracking as annotations of a target object are only available in the first frame of a test sequence. In this paper, we propose to learn hierarchical features for visual object tracking by using tree structure based Recursive Neural Networks (RNN), which have a relatively small number of parameters compared to other deep neural networks (e.g. Convolutional Neural Networks (CNN)) due to all basic modules in RNN share only one set of parameters. Experimental results demonstrate that our feature learning algorithm can significantly improve tracking performance on benchmark datasets.

Recent years, Discriminant Correlation Filter(DCF) has shown great advantages in the field of visual tracking, however its potential was greatly limited due to using single-resolution feature maps when applied to video with background interference. Therefore, this paper firstly used visual saliency to eliminate extra background information and outstand the object, and then extracted several features from different resolution images and merged them into a new feature vector. Then, the feature is used to train correlation filter templates for tracking. Compared with traditional algorithms, the proposed algorithm performed well on 51 benchmark videos of OTB. The method was robust to against challenges such as lighting changes, scale changes, occlusion, motion blur and while running at hundred frames-per-second, and was superior to other algorithms in distance accuracy and success rate.

In past years, deep convolutional neural networks (DCNN) have achieved big successes in image classification and object detection, as demonstrated on ImageNet in academic field. However, There are some unique practical challenges remain for real-world image recognition applications, e.g., small size of the objects, imbalanced data distributions, limited labeled data samples, etc. In this work, we are making efforts to deal with these challenges through a computational framework by incorporating latest developments in deep learning. In terms of two-stage detection scheme, pseudo labeling, data augmentation, cross-validation and ensemble learning, the proposed framework aims to achieve better performances for practical image recognition applications as compared to using standard deep learning methods. The proposed framework has recently been deployed as the key kernel for several image recognition competitions organized by Kaggle. The performance is promising as our final private scores were ranked 4 out of 2293 teams for fish recognition on the challenge "The Nature Conservancy Fisheries Monitoring" and 3 out of 834 teams for cervix recognition on the challenge "Intel &MobileODT Cervical Cancer Screening", and several others. We believe that by sharing the solutions, we can further promote the applications of deep learning techniques.

Automatic skin lesion analysis involves two critical steps: lesion segmentation and lesion classification. In this work, we propose a novel multi-target deep convolutional neural network (DCNN) to simultaneously tackle the problem of segmentation and classification. Based on U-Net and GoogleNet, a single model is constructed with three different targets of both lesion segmentation and two independent binary lesion classifications (i.e., melanoma detection and seborrheic keratosis identification), aiming to explore the differences and commonalities over different target models. We conduct experiments on dermoscopic images from the International Skin Imaging Collaboration (ISIC) 2017 Challenge. Results of our multi-target DCNN model demonstrates superiority over single model with one target only (such as U-net or GoogleNet), indicating its learning efficiency and potential for application in automatic skin lesion diagnosis. To the best of our knowledge, this work is the first demonstration for a single end-to-end deep neural network model that simultaneously handle both segmentation and classification in the field of skin lesion analysis.

In this paper, we study the challenging problem of multi-object tracking in a complex scene captured by a single camera. Different from the existing tracklet association-based tracking methods, we propose a novel and efficient way to obtain discriminative appearance-based tracklet affinity models. Our proposed method jointly learns the convolutional neural networks (CNNs) and temporally constrained metrics. In our method, a siamese convolutional neural network (CNN) is first pre-trained on the auxiliary data. Then the siamese CNN and temporally constrained metrics are jointly learned online to construct the appearance-based tracklet affinity models. The proposed method can jointly learn the hierarchical deep features and temporally constrained segment-wise metrics under a unified framework. For reliable association between tracklets, a novel loss function incorporating temporally constrained multi-task learning mechanism is proposed. By employing the proposed method, tracklet association can be accomplished even in challenging situations. Moreover, a largescale dataset with 40 fully annotated sequences is created to facilitate the tracking evaluation. Experimental results on five public datasets and the new large-scale dataset show that our method outperforms several state-of-the-art approaches in multi- object tracking.

In order to better deal with the partial occlusion issue, part-based trackers are widely used in visual object tracking recently. However, it is still difficult to realize fast and robust tracking, due to complicated online training and updating process. Correlation filters have been used in visual object tracking tasks recently because of their high efficiency. However, the traditional correlation filter based tracking methods do not deal with occlusion well. In this paper, we propose a novel tracking method which tracks objects based on parts with multiple correlation filters. The Bayesian inference framework and a structural constraint mask are adopted to enable our tracker to be robust to various appearance changes. Additionally, a discriminative part selection scheme is adopted to further improve performance and accelerate our method. Experimental results demonstrate that our multiple part tracker can significantly improve tracking performance on benchmark datasets.

In this paper, we present a novel method based on online target-specific metric learning and coherent dynamics estimation for tracklet (track fragment) association by network flow optimization in long-term multi-person tracking. Our proposed framework aims to exploit appearance and motion cues to prevent identity switches during tracking and to recover missed detections. Furthermore, target-specific metrics (appearance cue) and motion dynamics (motion cue) are proposed to be learned and estimated online, i.e., during the tracking process. Our approach is effective even when such cues fail to identify or follow the target due to occlusions or object-to-object interactions. We also propose to learn the weights of these two tracking cues to handle the difficult situations, such as severe occlusions and object-to-object interactions effectively. Our method has been validated on several public datasets and the experimental results show that it outperforms several state-of-the-art tracking methods.