Domain adaptive detection aims to improve the generalization of detectors on target domain. To reduce discrepancy in feature distributions between two domains, recent approaches achieve domain adaption through feature alignment in different granularities via adversarial learning. However, they neglect the relationship between multiple granularities and different features in alignment, degrading detection. Addressing this, we introduce a unified multi-granularity alignment (MGA)-based detection framework for domain-invariant feature learning. The key is to encode the dependencies across different granularities including pixel-, instance-, and category-levels simultaneously to align two domains. Specifically, based on pixel-level features, we first develop an omni-scale gated fusion (OSGF) module to aggregate discriminative representations of instances with scale-aware convolutions, leading to robust multi-scale detection. Besides, we introduce multi-granularity discriminators to identify where, either source or target domains, different granularities of samples come from. Note that, MGA not only leverages instance discriminability in different categories but also exploits category consistency between two domains for detection. Furthermore, we present an adaptive exponential moving average (AEMA) strategy that explores model assessments for model update to improve pseudo labels and alleviate local misalignment problem, boosting detection robustness. Extensive experiments on multiple domain adaption scenarios validate the superiority of MGA over other approaches on FCOS and Faster R-CNN detectors. Code will be released at https://github.com/tiankongzhang/MGA.

图像介入寻求一种语义一致的方法，以根据其未掩盖的内容来恢复损坏的图像。以前的方法通常将训练有素的甘恩重复使用，然后在产生逼真的斑块中用于缺少GAN反转的孔。然而，在这些算法中对硬约束的无知可能会产生gan倒置和图像插入之间的差距。在解决这个问题的情况下，我们在本文中设计了一个新颖的GAN反转模型，用于图像插入，称为Interverfill，主要由带有预调制模块的编码器和具有F＆W+潜在空间的GAN生成器组成。在编码器中，预调制网络利用多尺度结构将更多的歧视语义编码为样式向量。为了弥合GAN倒置和图像插入之间的缝隙，提出了F＆W+潜在空间以消除巨大的颜色差异和语义不一致。为了重建忠实和逼真的图像，一个简单而有效的软上升平均潜在模块旨在捕获更多样化的内域模式，以合成大型腐败的高保真质地。在包括Ploce2，Celeba-HQ，Metfaces和Scenery在内的四个具有挑战性的数据集上进行的全面实验表明，我们的Intervill效果优于定性和定量的高级方法，并支持室外图像的完成。

依靠变压器进行复杂的视觉功能学习，对象跟踪目睹了最先进的新标准（SOTA）。但是，这种进步伴随着更大的培训数据和更长的培训期，使跟踪越来越昂贵。在本文中，我们证明了变压器的依赖性不是必需的，并且在实现SOTA跟踪方面，纯粹的convnets仍然具有竞争力，甚至更经济和友好。我们的解决方案是释放多模式视觉语言（VL）跟踪的功能，只需使用Convnet。本质在于通过我们的模态混音器（Modamixer）和不对称Convnet搜索学习新颖的统一自适应VL表示。我们表明，我们的统一自适应VL表示形式纯粹是用Convnet学习的，是变压器视觉特征的简单而强大的替代方案，它令人难以置信地将基于CNN的基于CNN的Siamese Tracker提高了14.5％的SUC，在挑战性的Lasot方面（50.7％> 65.2％> 65.2％> 65.2％ ），即使表现优于几个基于变压器的SOTA跟踪器。除经验结果外，我们理论上分析了我们的方法以证明其有效性。通过揭示VL代表的潜力，我们希望社区将更多的关注转移到VL跟踪上，并希望为超越变形金刚的未来跟踪开放更多的可能性。代码和模型将在https://github.com/judasdie/sots上发布。

Multi-animal tracking (MAT), a multi-object tracking (MOT) problem, is crucial for animal motion and behavior analysis and has many crucial applications such as biology, ecology and animal conservation. Despite its importance, MAT is largely under-explored compared to other MOT problems such as multi-human tracking due to the scarcity of dedicated benchmarks. To address this problem, we introduce AnimalTrack, a dedicated benchmark for multi-animal tracking in the wild. Specifically, AnimalTrack consists of 58 sequences from a diverse selection of 10 common animal categories. On average, each sequence comprises of 33 target objects for tracking. In order to ensure high quality, every frame in AnimalTrack is manually labeled with careful inspection and refinement. To our best knowledge, AnimalTrack is the first benchmark dedicated to multi-animal tracking. In addition, to understand how existing MOT algorithms perform on AnimalTrack and provide baselines for future comparison, we extensively evaluate 14 state-of-the-art representative trackers. The evaluation results demonstrate that, not surprisingly, most of these trackers become degenerated due to the differences between pedestrians and animals in various aspects (e.g., pose, motion, and appearance), and more efforts are desired to improve multi-animal tracking. We hope that AnimalTrack together with evaluation and analysis will foster further progress on multi-animal tracking. The dataset and evaluation as well as our analysis will be made available at https://hengfan2010.github.io/projects/AnimalTrack/.

图像分割是医学图像场中的重要任务，并且已经提出了许多基于卷积神经网络（CNNS）的方法，其中U-Net及其变体表现出了有希望的性能。在本文中，我们提出了基于U-Net的GP模块和GPU-Net，通过引入幽灵模块和不足的空间金字塔池（ASPP），可以了解更多样化的功能。我们的方法实现了更好的性能，参数较少的4倍以上，拖鞋的2倍，为未来的研究提供了新的潜在方向。我们的即插即用模块也可以应用于现有的分段方法，以进一步提高其性能。

我们介绍了一种新颖的骨干架构，提高特征表示的目标感知能力。具体地，已经观察到事实上框架简单地使用来自骨干网的输出来执行特征匹配，从备份目标本地化，没有从匹配模块到骨干网的直接反馈，尤其是浅层。更具体地，只有匹配模块可以直接访问目标信息（在参考帧中），而候选帧的表示学习对参考目标是盲目的。结果，浅级中的目标 - 无关干扰的累积效果可能降低更深层的特征质量。在本文中，我们通过在暹罗类似的骨干网（inbn）内进行多个分支 - 方面交互来从不同角度接近问题。在INBN的核心是一个通用交互建模器（GIM），其将参考图像的先前知识注入骨干网络的不同阶段，导致候选特征表示的更好的目标感知和鲁棒的牵引力，其计算成本具有可忽略的计算成本。所提出的GIM模块和INBN机制是一般的，适用于不同的骨干类型，包括CNN和变压器，以改进，如我们在多个基准上的广泛实验所证明的那样。特别是，CNN版本（基于Siamcar），分别在Lasot / TNL2K上改善了3.2 / 6.9的Suc绝对收益。变压器版本获取Lasot / TNL2K的SUC 25.7 / 52.0，与最近的艺术态度相提并论。代码和模型将被释放。

变压器最近展示了改进视觉跟踪算法的明显潜力。尽管如此，基于变压器的跟踪器主要使用变压器熔断并增强由卷积神经网络（CNNS）产生的功能。相比之下，在本文中，我们提出了一个完全基于注意力的变压器跟踪算法，Swin-Cranstormer Tracker（SwintRack）。 SwintRack使用变压器进行特征提取和特征融合，允许目标对象和搜索区域之间的完全交互进行跟踪。为了进一步提高性能，我们调查了全面的不同策略，用于特征融合，位置编码和培训损失。所有这些努力都使SwintRack成为一个简单但坚实的基线。在我们的彻底实验中，SwintRack在leasot上设置了一个新的记录，在4.6 \％的情况下超过4.6 \％，同时仍然以45 fps运行。此外，它达到了最先进的表演，0.483 Suc，0.832 Suc和0.694 Ao，其他具有挑战性的leasot _ {ext} $，trackingnet和got-10k。我们的实施和培训型号可在HTTPS://github.com/litinglin/swintrack获得。

我们介绍了一个开源深学习库的Pytorchvideo，为各种视频理解任务提供了丰富的模块化，高效，可重复的组件，包括分类，检测，自我监督学习和低级处理。该库涵盖了一系列视频理解工具，包括复制最先进的性能的多模式数据加载，转换和模型。Pytorchvideo进一步支持硬件加速，从而实现移动设备上的实时推断。图书馆基于Pytorch，可以由任何培训框架使用;例如，pytorchlightning，pyslowfast或优雅的愿景。pytorchvideo在https://pytorchvideo.org/提供

Zero-Shot Learning has been a highlighted research topic in both vision and language areas. Recently, most existing methods adopt structured knowledge information to model explicit correlations among categories and use deep graph convolutional network to propagate information between different categories. However, it is difficult to add new categories to existing structured knowledge graph, and deep graph convolutional network suffers from over-smoothing problem. In this paper, we provide a new semantic enhanced knowledge graph that contains both expert knowledge and categories semantic correlation. Our semantic enhanced knowledge graph can further enhance the correlations among categories and make it easy to absorb new categories. To propagate information on the knowledge graph, we propose a novel Residual Graph Convolutional Network (ResGCN), which can effectively alleviate the problem of over-smoothing. Experiments conducted on the widely used large-scale ImageNet-21K dataset and AWA2 dataset show the effectiveness of our method, and establish a new state-of-the-art on zero-shot learning. Moreover, our results on the large-scale ImageNet-21K with various feature extraction networks show that our method has better generalization and robustness.

In subcellular biological research, fluorescence staining is a key technique to reveal the locations and morphology of subcellular structures. However, fluorescence staining is slow, expensive, and harmful to cells. In this paper, we treat it as a deep learning task termed subcellular structure prediction (SSP), aiming to predict the 3D fluorescent images of multiple subcellular structures from a 3D transmitted-light image. Unfortunately, due to the limitations of current biotechnology, each image is partially labeled in SSP. Besides, naturally, the subcellular structures vary considerably in size, which causes the multi-scale issue in SSP. However, traditional solutions can not address SSP well since they organize network parameters inefficiently and inflexibly. To overcome these challenges, we propose Re-parameterizing Mixture-of-Diverse-Experts (RepMode), a network that dynamically organizes its parameters with task-aware priors to handle specified single-label prediction tasks of SSP. In RepMode, the Mixture-of-Diverse-Experts (MoDE) block is designed to learn the generalized parameters for all tasks, and gating re-parameterization (GatRep) is performed to generate the specialized parameters for each task, by which RepMode can maintain a compact practical topology exactly like a plain network, and meanwhile achieves a powerful theoretical topology. Comprehensive experiments show that RepMode outperforms existing methods on ten of twelve prediction tasks of SSP and achieves state-of-the-art overall performance.