Dong Chen

Yixuan Wei, Han Hu, Zhenda Xie et al.

Masked image modeling (MIM) learns representations with remarkably good fine-tuning performances, overshadowing previous prevalent pre-training approaches such as image classification, instance contrastive learning, and image-text alignment. In this paper, we show that the inferior fine-tuning performance of these pre-training approaches can be significantly improved by a simple post-processing in the form of feature distillation (FD). The feature distillation converts the old representations to new representations that have a few desirable properties just like those representations produced by MIM. These properties, which we aggregately refer to as optimization friendliness, are identified and analyzed by a set of attention- and optimization-related diagnosis tools. With these properties, the new representations show strong fine-tuning performance. Specifically, the contrastive self-supervised learning methods are made as competitive in fine-tuning as the state-of-the-art masked image modeling (MIM) algorithms. The CLIP models' fine-tuning performance is also significantly improved, with a CLIP ViT-L model reaching 89.0% top-1 accuracy on ImageNet-1K classification. On the 3-billion-parameter SwinV2-G model, the fine-tuning accuracy is improved by +1.5 mIoU / +1.1 mAP to 61.4 mIoU / 64.2 mAP on ADE20K semantic segmentation and COCO object detection, respectively, creating new records on both benchmarks. More importantly, our work provides a way for the future research to focus more effort on the generality and scalability of the learnt representations without being pre-occupied with optimization friendliness since it can be enhanced rather easily. The code will be available at https://github.com/SwinTransformer/Feature-Distillation.

Xiaoyi Dong, Jianmin Bao, Ting Zhang et al.

We propose bootstrapped masked autoencoders (BootMAE), a new approach for vision BERT pretraining. BootMAE improves the original masked autoencoders (MAE) with two core designs: 1) momentum encoder that provides online feature as extra BERT prediction targets; 2) target-aware decoder that tries to reduce the pressure on the encoder to memorize target-specific information in BERT pretraining. The first design is motivated by the observation that using a pretrained MAE to extract the features as the BERT prediction target for masked tokens can achieve better pretraining performance. Therefore, we add a momentum encoder in parallel with the original MAE encoder, which bootstraps the pretraining performance by using its own representation as the BERT prediction target. In the second design, we introduce target-specific information (e.g., pixel values of unmasked patches) from the encoder directly to the decoder to reduce the pressure on the encoder of memorizing the target-specific information. Thus, the encoder focuses on semantic modeling, which is the goal of BERT pretraining, and does not need to waste its capacity in memorizing the information of unmasked tokens related to the prediction target. Through extensive experiments, our BootMAE achieves $84.2\%$ Top-1 accuracy on ImageNet-1K with ViT-B backbone, outperforming MAE by $+0.8\%$ under the same pre-training epochs. BootMAE also gets $+1.0$ mIoU improvements on semantic segmentation on ADE20K and $+1.3$ box AP, $+1.4$ mask AP improvement on object detection and segmentation on COCO dataset. Code is released at https://github.com/LightDXY/BootMAE.

Hanqing Zhao, Dianmo Sheng, Jianmin Bao et al.

Copy-Paste is a simple and effective data augmentation strategy for instance segmentation. By randomly pasting object instances onto new background images, it creates new training data for free and significantly boosts the segmentation performance, especially for rare object categories. Although diverse, high-quality object instances used in Copy-Paste result in more performance gain, previous works utilize object instances either from human-annotated instance segmentation datasets or rendered from 3D object models, and both approaches are too expensive to scale up to obtain good diversity. In this paper, we revisit Copy-Paste at scale with the power of newly emerged zero-shot recognition models (e.g., CLIP) and text2image models (e.g., StableDiffusion). We demonstrate for the first time that using a text2image model to generate images or zero-shot recognition model to filter noisily crawled images for different object categories is a feasible way to make Copy-Paste truly scalable. To make such success happen, we design a data acquisition and processing framework, dubbed ``X-Paste", upon which a systematic study is conducted. On the LVIS dataset, X-Paste provides impressive improvements over the strong baseline CenterNet2 with Swin-L as the backbone. Specifically, it archives +2.6 box AP and +2.1 mask AP gains on all classes and even more significant gains with +6.8 box AP, +6.5 mask AP on long-tail classes. Our code and models are available at https://github.com/yoctta/XPaste.

Dong Chen, Huili Peng, Yong Hu et al.

Objective: To systematically quantify the effect of the camera view (frontal vs. lateral) on the accuracy of 2D markerless gait analysis relative to 3D motion capture ground truth. Methods: Gait data from 18 subjects were recorded simultaneously using frontal, lateral and 3D motion capture systems. Pose estimation used YOLOv8. Four metrics were assessed to evaluate agreement: Dynamic Time Warping (DTW) for temporal alignment, Maximum Cross-Correlation (MCC) for signal similarity, Kullback-Leibler Divergence (KLD) for distribution differences, and Information Entropy (IE) for complexity. Wilcoxon signed-rank tests (significance: $p < 0.05$) and Cliff's delta ($δ$) were used to measure statistical differences and effect sizes. Results: Lateral views significantly outperformed frontal views for sagittal plane kinematics: step length (DTW: $53.08 \pm 24.50$ vs. $69.87 \pm 25.36$, $p = 0.005$) and knee rotation (DTW: $106.46 \pm 38.57$ vs. $155.41 \pm 41.77$, $p = 0.004$). Frontal views were superior for symmetry parameters: trunk rotation (KLD: $0.09 \pm 0.06$ vs. $0.30 \pm 0.19$, $p < 0.001$) and wrist-to-hipmid distance (MCC: $105.77 \pm 29.72$ vs. $75.20 \pm 20.38$, $p = 0.003$). Effect sizes were medium-to-large ($δ: 0.34$--$0.76$). Conclusion: Camera view critically impacts gait parameter accuracy. Lateral views are optimal for sagittal kinematics; frontal views excel for trunk symmetry. Significance: This first systematic evidence enables data-driven camera deployment in 2D gait analysis, enhancing clinical utility. Future implementations should leverage both views via disease-oriented setups.

Wangzheng Shi, Yinglin Zheng, Yuxin Lin et al.

Hair transfer is increasingly valuable across domains such as social media, gaming, advertising, and entertainment. While significant progress has been made in single-image hair transfer, video-based hair transfer remains challenging due to the need for temporal consistency, spatial fidelity, and dynamic adaptability. In this work, we propose HairShifter, a novel "Anchor Frame + Animation" framework that unifies high-quality image hair transfer with smooth and coherent video animation. At its core, HairShifter integrates a Image Hair Transfer (IHT) module for precise per-frame transformation and a Multi-Scale Gated SPADE Decoder to ensure seamless spatial blending and temporal coherence. Our method maintains hairstyle fidelity across frames while preserving non-hair regions. Extensive experiments demonstrate that HairShifter achieves state-of-the-art performance in video hairstyle transfer, combining superior visual quality, temporal consistency, and scalability. The code will be publicly available. We believe this work will open new avenues for video-based hairstyle transfer and establish a robust baseline in this field.

Zhenyue Wang, Guowu Yuan, Hao Zhou et al.

Insulators are crucial insulation components and structural supports in power grids, playing a vital role in the transmission lines. Due to temperature fluctuations, internal stress, or damage from hail, insulators are prone to injury. Automatic detection of damaged insulators faces challenges such as diverse types, small defect targets, and complex backgrounds and shapes. Most research for detecting insulator defects has focused on a single defect type or a specific material. However, the insulators in the grid's transmission lines have different colors and materials. Various insulator defects coexist, and the existing methods have difficulty meeting the practical application requirements. Current methods suffer from low detection accuracy and mAP0.5 cannot meet application requirements. This paper proposes an improved YOLOv7 model for multi-type insulator defect detection. First, our model replaces the SPPCSPC module with the RFB module to enhance the network's feature extraction capability. Second, a CA mechanism is introduced into the head part to enhance the network's feature representation ability and to improve detection accuracy. Third, a WIoU loss function is employed to address the low-quality samples hindering model generalization during training, thereby improving the model's overall performance. The experimental results indicate that the proposed model exhibits enhancements across various performance metrics. Specifically, there is a 1.6% advancement in mAP_0.5, a corresponding 1.6% enhancement in mAP_0.5:0.95, a 1.3% elevation in precision, and a 1% increase in recall. Moreover, the model achieves parameter reduction by 3.2 million, leading to a decrease of 2.5 GFLOPS in computational cost. Notably, there is also an improvement of 2.81 milliseconds in single-image detection speed.

Xiaoyi Dong, Jianmin Bao, Ting Zhang et al.

This paper explores a better prediction target for BERT pre-training of vision transformers. We observe that current prediction targets disagree with human perception judgment.This contradiction motivates us to learn a perceptual prediction target. We argue that perceptually similar images should stay close to each other in the prediction target space. We surprisingly find one simple yet effective idea: enforcing perceptual similarity during the dVAE training. Moreover, we adopt a self-supervised transformer model for deep feature extraction and show that it works well for calculating perceptual similarity.We demonstrate that such learned visual tokens indeed exhibit better semantic meanings, and help pre-training achieve superior transfer performance in various downstream tasks. For example, we achieve $\textbf{84.5\%}$ Top-1 accuracy on ImageNet-1K with ViT-B backbone, outperforming the competitive method BEiT by $\textbf{+1.3\%}$ under the same pre-training epochs. Our approach also gets significant improvement on object detection and segmentation on COCO and semantic segmentation on ADE20K. Equipped with a larger backbone ViT-H, we achieve the state-of-the-art ImageNet accuracy (\textbf{88.3\%}) among methods using only ImageNet-1K data.

Dong Chen, Duoqian Miao

Numerous improvements for feedback mechanisms have contributed to the great progress in object detection. In this paper, we first present an evaluation-feedback module, which is proposed to consist of evaluation system and feedback mechanism. Then we analyze and summarize the disadvantages and improvements of traditional evaluation-feedback module. Finally, we focus on both the evaluation system and the feedback mechanism, and propose Control Distance IoU and Control Distance IoU loss function (or CDIoU and CDIoU loss for short) without increasing parameters or FLOPs in models, which show different significant enhancements on several classical and emerging models. Some experiments and comparative tests show that coordinated evaluation-feedback module can effectively improve model performance. CDIoU and CDIoU loss have different excellent performances in several models such as Faster R-CNN, YOLOv4, RetinaNet and ATSS. There is a maximum AP improvement of 1.9% and an average AP of 0.8% improvement on MS COCO dataset, compared to traditional evaluation-feedback modules.