Transformers have been successfully applied to the visual tracking task and significantly promote tracking performance. The self-attention mechanism designed to model long-range dependencies is the key to the success of Transformers. However, self-attention lacks focusing on the most relevant information in the search regions, making it easy to be distracted by background. In this paper, we relieve this issue with a sparse attention mechanism by focusing the most relevant information in the search regions, which enables a much accurate tracking. Furthermore, we introduce a double-head predictor to boost the accuracy of foreground-background classification and regression of target bounding boxes, which further improve the tracking performance. Extensive experiments show that, without bells and whistles, our method significantly outperforms the state-of-the-art approaches on LaSOT, GOT-10k, TrackingNet, and UAV123, while running at 40 FPS. Notably, the training time of our method is reduced by 75% compared to that of TransT. The source code and models are available at https://github.com/fzh0917/SparseTT.
Pan-sharpening aims at producing a high-resolution (HR) multi-spectral (MS) image from a low-resolution (LR) multi-spectral (MS) image and its corresponding panchromatic (PAN) image acquired by a same satellite. Inspired by a new fashion in recent deep learning community, we propose a novel Transformer based model for pan-sharpening. We explore the potential of Transformer in image feature extraction and fusion. Following the successful development of vision transformers, we design a two-stream network with the self-attention to extract the modality-specific features from the PAN and MS modalities and apply a cross-attention module to merge the spectral and spatial features. The pan-sharpened image is produced from the enhanced fused features. Extensive experiments on GaoFen-2 and WorldView-3 images demonstrate that our Transformer based model achieves impressive results and outperforms many existing CNN based methods, which shows the great potential of introducing Transformer to the pan-sharpening task. Codes are available at https://github.com/zhysora/PanFormer.
Human gait is considered a unique biometric identifier which can be acquired in a covert manner at a distance. However, models trained on existing public domain gait datasets which are captured in controlled scenarios lead to drastic performance decline when applied to real-world unconstrained gait data. On the other hand, video person re-identification techniques have achieved promising performance on large-scale publicly available datasets. Given the diversity of clothing characteristics, clothing cue is not reliable for person recognition in general. So, it is actually not clear why the state-of-the-art person re-identification methods work as well as they do. In this paper, we construct a new gait dataset by extracting silhouettes from an existing video person re-identification challenge which consists of 1,404 persons walking in an unconstrained manner. Based on this dataset, a consistent and comparative study between gait recognition and person re-identification can be carried out. Given that our experimental results show that current gait recognition approaches designed under data collected in controlled scenarios are inappropriate for real surveillance scenarios, we propose a novel gait recognition method, called RealGait. Our results suggest that recognizing people by their gait in real surveillance scenarios is feasible and the underlying gait pattern is probably the true reason why video person re-idenfification works in practice.
Occlusions are very common in face images in the wild, leading to the degraded performance of face-related tasks. Although much effort has been devoted to removing occlusions from face images, the varying shapes and textures of occlusions still challenge the robustness of current methods. As a result, current methods either rely on manual occlusion masks or only apply to specific occlusions. This paper proposes a novel face de-occlusion model based on face segmentation and 3D face reconstruction, which automatically removes all kinds of face occlusions with even blurred boundaries,e.g., hairs. The proposed model consists of a 3D face reconstruction module, a face segmentation module, and an image generation module. With the face prior and the occlusion mask predicted by the first two, respectively, the image generation module can faithfully recover the missing facial textures. To supervise the training, we further build a large occlusion dataset, with both manually labeled and synthetic occlusions. Qualitative and quantitative results demonstrate the effectiveness and robustness of the proposed method.
Dance challenges are going viral in video communities like TikTok nowadays. Once a challenge becomes popular, thousands of short-form videos will be uploaded in merely a couple of days. Therefore, virality prediction from dance challenges is of great commercial value and has a wide range of applications, such as smart recommendation and popularity promotion. In this paper, a novel multi-modal framework which integrates skeletal, holistic appearance, facial and scenic cues is proposed for comprehensive dance virality prediction. To model body movements, we propose a pyramidal skeleton graph convolutional network (PSGCN) which hierarchically refines spatio-temporal skeleton graphs. Meanwhile, we introduce a relational temporal convolutional network (RTCN) to exploit appearance dynamics with non-local temporal relations. An attentive fusion approach is finally proposed to adaptively aggregate predictions from different modalities. To validate our method, we introduce a large-scale viral dance video (VDV) dataset, which contains over 4,000 dance clips of eight viral dance challenges. Extensive experiments on the VDV dataset demonstrate the efficacy of our model. Extensive experiments on the VDV dataset well demonstrate the effectiveness of our approach. Furthermore, we show that short video applications like multi-dimensional recommendation and action feedback can be derived from our model.
Though image-level weakly supervised semantic segmentation (WSSS) has achieved great progress with Class Activation Map (CAM) as the cornerstone, the large supervision gap between classification and segmentation still hampers the model to generate more complete and precise pseudo masks for segmentation. In this study, we explore two implicit but intuitive constraints, i.e., cross-view feature semantic consistency and intra(inter)-class compactness(dispersion), to narrow the supervision gap. To this end, we propose two novel pixel-to-prototype contrast regularization terms that are conducted cross different views and within per single view of an image, respectively. Besides, we adopt two sample mining strategies, named semi-hard prototype mining and hard pixel sampling, to better leverage hard examples while reducing incorrect contrasts caused due to the absence of precise pixel-wise labels. Our method can be seamlessly incorporated into existing WSSS models without any changes to the base network and does not incur any extra inference burden. Experiments on standard benchmark show that our method consistently improves two strong baselines by large margins, demonstrating the effectiveness of our method. Specifically, built on top of SEAM, we improve the initial seed mIoU on PASCAL VOC 2012 from 55.4% to 61.5%. Moreover, armed with our method, we increase the segmentation mIoU of EPS from 70.8% to 73.6%, achieving new state-of-the-art.
Deep learning based pan-sharpening has received significant research interest in recent years. Most of existing methods fall into the supervised learning framework in which they down-sample the multi-spectral (MS) and panchromatic (PAN) images and regard the original MS images as ground truths to form training samples. Although impressive performance could be achieved, they have difficulties generalizing to the original full-scale images due to the scale gap, which makes them lack of practicability. In this paper, we propose an unsupervised generative adversarial framework that learns from the full-scale images without the ground truths to alleviate this problem. We extract the modality-specific features from the PAN and MS images with a two-stream generator, perform fusion in the feature domain, and then reconstruct the pan-sharpened images. Furthermore, we introduce a novel hybrid loss based on the cycle-consistency and adversarial scheme to improve the performance. Comparison experiments with the state-of-the-art methods are conducted on GaoFen-2 and WorldView-3 satellites. Results demonstrate that the proposed method can greatly improve the pan-sharpening performance on the full-scale images, which clearly show its practical value. Codes and datasets will be made publicly available.
Differentiable ARchiTecture Search (DARTS) uses a continuous relaxation of network representation and dramatically accelerates Neural Architecture Search (NAS) by almost thousands of times in GPU-day. However, the searching process of DARTS is unstable, which suffers severe degradation when training epochs become large, thus limiting its application. In this paper, we claim that this degradation issue is caused by the imbalanced norms between different nodes and the highly correlated outputs from various operations. We then propose an improved version of DARTS, namely iDARTS, to deal with the two problems. In the training phase, it introduces node normalization to maintain the norm balance. In the discretization phase, the continuous architecture is approximated based on the similarity between the outputs of the node and the decorrelated operations rather than the values of the architecture parameters. Extensive evaluation is conducted on CIFAR-10 and ImageNet, and the error rates of 2.25\% and 24.7\% are reported within 0.2 and 1.9 GPU-day for architecture search respectively, which shows its effectiveness. Additional analysis also reveals that iDARTS has the advantage in robustness and generalization over other DARTS-based counterparts.
The transferability and robustness of adversarial examples are two practical yet important properties for black-box adversarial attacks. In this paper, we explore effective mechanisms to boost both of them from the perspective of network hierarchy, where a typical network can be hierarchically divided into output stage, intermediate stage and input stage. Since over-specialization of source model, we can hardly improve the transferability and robustness of the adversarial perturbations in the output stage. Therefore, we focus on the intermediate and input stages in this paper and propose a transferable and robust adversarial perturbation generation (TRAP) method. Specifically, we propose the dynamically guided mechanism to continuously calculate accurate directional guidances for perturbation generation in the intermediate stage. In the input stage, instead of the single-form transformation augmentations adopted in the existing methods, we leverage multiform affine transformation augmentations to further enrich the input diversity and boost the robustness and transferability of the adversarial perturbations. Extensive experiments demonstrate that our TRAP achieves impressive transferability and high robustness against certain interferences.