Recent advancements in diffusion models have significantly enhanced the performance of text-to-image models in image synthesis. To enable control over the the spatial locations of the generated objects, diffusion-based methods typically utilize object layout as an auxiliary input. However, we observe that this approach treats all objects as being on the same layer and neglect their visibility order, leading to the synthesis of overlapping objects with incorrect occlusions. To address this limitation, we introduce in this paper a new training-free framework that considers object visibility order explicitly and allows users to place overlapping objects in a stack of layers. Our framework consists of two visibility-based designs. First, we propose a novel Sequential Denoising Process (SDP) to divide the whole image generation into multiple stages for different objects, each stage primarily focuses on an object. Second, we propose a novel Visibility-Order-Aware (VOA) Loss to transform the layout and occlusion constraints into an attention map optimization process to improve the accuracy of synthesizing object occlusions in complex scenes. By merging these two novel components, our framework, dubbed VODiff, enables the generation of photorealistic images that satisfy user-specified spatial constraints and object occlusion relationships. In addition, we introduce VOBench, a diverse benchmark dataset containing 200 curated samples, each with a reference image, text prompts, object visibility orders and layout maps. We conduct extensive evaluations on this dataset to demonstrate the superiority of our approach.

Salient Object Ranking (SOR) aims to study human attention shifts across different objects in the scene. It is a challenging task, as it requires comprehension of the relations among the salient objects in the scene. However, existing works often overlook such relations or model them implicitly. In this work, we observe that when Large Vision-Language Models (LVLMs) describe a scene, they usually focus on the most salient object first, and then discuss the relations as they move on to the next (less salient) one. Based on this observation, we propose a novel Language-Guided Salient Object Ranking approach (named LG-SOR), which utilizes the internal knowledge within the LVLM-generated language descriptions, i.e., semantic relation cues and the implicit entity order cues, to facilitate saliency ranking. Specifically, we first propose a novel Text-Guided Visual Modulation (TGVM) module to incorporate semantic information in the description for saliency ranking. TGVM controls the flow of linguistic information to the visual features, suppresses noisy background image features, and enables the propagation of useful textual features. We then propose a novel Text-Aware Visual Reasoning (TAVR) module to enhance model reasoning in object ranking, by explicitly learning a multimodal graph based on the entity and relation cues derived from the description. Extensive experiments demonstrate superior performances of our model on two SOR benchmarks.

While existing image‑guided composition methods may help insert a foreground object onto a user-specified region of a background image, achieving natural blending inside the region with the rest of the image unchanged, we observe that these existing methods often struggle in synthesizing seamless interaction-aware compositions when the task involves human-object interactions. In this paper, we first propose HOComp, a novel approach for compositing a foreground object onto a human-centric background image, while ensuring harmonious interactions between the foreground object and the background person and their consistent appearances. Our approach includes two key designs: (1) MLLMs-driven Region-based Pose Guidance (MRPG), which utilizes MLLMs to identify the interaction region as well as the interaction type (e.g., holding and lefting) to provide coarse-to-fine constraints to the generated pose for the interaction while incorporating human pose landmarks to track action variations and enforcing fine-grained pose constraints; and (2) Detail-Consistent Appearance Preservation (DCAP), which unifies a shape-aware attention modulation mechanism, a multi-view appearance loss, and a background consistency loss to ensure consistent shapes/textures of the foreground and faithful reproduction of the background human. We then propose the first dataset, named Interaction-aware Human-Object Composition (IHOC), for the task. Experimental results on our dataset show that HOComp effectively generates harmonious human-object interactions with consistent appearances, and outperforms relevant methods qualitatively and quantitatively.

Spatio-temporal video grounding (STVG) aims at localizing the spatio-temporal tube of a video, as specified by the input text query. In this paper, we utilize multimodal large language models (MLLMs) to explore a zero-shot solution in STVG. We reveal two key insights about MLLMs: (1) MLLMs tend to dynamically assign special tokens, referred to as \textit{grounding tokens}, for grounding the text query; and (2) MLLMs often suffer from suboptimal grounding due to the inability to fully integrate the cues in the text query (\textit{e.g.}, attributes, actions) for inference. Based on these insights, we propose a MLLM-based zero-shot framework for STVG, which includes novel decomposed spatio-temporal highlighting (DSTH) and temporal-augmented assembling (TAS) strategies to unleash the reasoning ability of MLLMs. The DSTH strategy first decouples the original query into attribute and action sub-queries for inquiring the existence of the target both spatially and temporally. It then uses a novel logit-guided re-attention (LRA) module to learn latent variables as spatial and temporal prompts, by regularizing token predictions for each sub-query. These prompts highlight attribute and action cues, respectively, directing the model's attention to reliable spatial and temporal related visual regions. In addition, as the spatial grounding by the attribute sub-query should be temporally consistent, we introduce the TAS strategy to assemble the predictions using the original video frames and the temporal-augmented frames as inputs to help improve temporal consistency. We evaluate our method on various MLLMs, and show that it outperforms SOTA methods on three common STVG benchmarks.

Recent advances in deep generative modeling have unlocked unprecedented opportunities for video synthesis. In real-world applications, however, users often seek tools to faithfully realize their creative editing intentions with precise and consistent control. Despite the progress achieved by existing methods, ensuring fine-grained alignment with user intentions remains an open and challenging problem. In this work, we present Shape-for-Motion, a novel framework that incorporates a 3D proxy for precise and consistent video editing. Shape-for-Motion achieves this by converting the target object in the input video to a time-consistent mesh, i.e., a 3D proxy, allowing edits to be performed directly on the proxy and then inferred back to the video frames. To simplify the editing process, we design a novel Dual-Propagation Strategy that allows users to perform edits on the 3D mesh of a single frame, and the edits are then automatically propagated to the 3D meshes of the other frames. The 3D meshes for different frames are further projected onto the 2D space to produce the edited geometry and texture renderings, which serve as inputs to a decoupled video diffusion model for generating edited results. Our framework supports various precise and physically-consistent manipulations across the video frames, including pose editing, rotation, scaling, translation, texture modification, and object composition. Our approach marks a key step toward high-quality, controllable video editing workflows. Extensive experiments demonstrate the superiority and effectiveness of our approach. Project Page: https://shapeformotion.github.io.

Real-world applications like video gaming and virtual reality often demand the ability to model 3D scenes that users can explore along custom camera trajectories. While significant progress has been made in generating 3D objects from text or images, creating long-range, 3D-consistent, explorable 3D scenes remains a complex and challenging problem. In this work, we present Voyager , a novel video diffusion framework that generates world-consistent 3D point-cloud sequences from a single image with user-defined camera path. Unlike existing approaches, Voyager achieves end-to-end scene generation and reconstruction with inherent consistency across frames, eliminating the need for 3D reconstruction pipelines (e.g., structure-from-motion or multi-view stereo). Our method integrates three key components: 1) World-Consistent Video Diffusion : A unified architecture that jointly generates aligned RGB and depth video sequences, conditioned on existing world observation to ensure global coherence 2) Long-Range World Exploration : An efficient world cache with point culling and an auto-regressive inference with smooth video sampling for iterative scene extension with context-aware consistency, and 3) Scalable Data Engine : A video reconstruction pipeline that automates camera pose estimation and metric depth prediction for arbitrary videos, enabling large-scale, diverse training data curation without manual 3D annotations. Collectively, these designs result in a clear improvement over existing methods in visual quality and geometric accuracy, with versatile applications. Code for this paper are at https://github.com/Tencent-Hunyuan/HunyuanWorld-Voyager.

Removing shadows requires an understanding of both lighting conditions and object textures in a scene. Existing methods typically learn pixel-level color mappings between shadow and non-shadow images, in which the joint modeling of lighting and object textures is implicit and inadequate. We observe that in a shadow region, the degradation degree of object textures depends on the local illumination, while simply enhancing the local illumination cannot fully recover the attenuated textures. Based on this observation, we propose to condition the restoration of attenuated textures on the corrected local lighting in the shadow region. Specifically, We first design a shadow-aware decomposition network to estimate the illumination and reflectance layers of shadow regions explicitly. We then propose a novel bilateral correction network to recast the lighting of shadow regions in the illumination layer via a novel local lighting correction module, and to restore the textures conditioned on the corrected illumination layer via a novel illumination-guided texture restoration module. We further annotate pixel-wise shadow masks for the public SRD dataset, which originally contains only image pairs. Experiments on three benchmarks show that our method outperforms existing state-of-the-art shadow removal methods. Project page in: yuhaoliu7456.github.io/RRL-Net.

Recent research has shown significant interest in image-based glass surface detection (GSD). However, detecting glass surfaces in dynamic scenes remains largely unexplored due to the lack of a high-quality dataset and an effective video glass surface detection (VGSD) method. In this paper, we propose the first VGSD approach. Our key observation is that reflections frequently appear on glass surfaces, but they change dynamically as the camera moves. Based on this observation, we propose to offset the excessive dependence on a single uncertainty reflection via joint modeling of temporal and spatial reflection cues. To this end, we propose the VGSD-Net with two novel modules: a Location-aware Reflection Extraction (LRE) module and a Context-enhanced Reflection Integration (CRI) module, for the position-aware reflection feature extraction and the spatial-temporal reflection cues integration, respectively. We have also created the first large-scale video glass surface dataset (VGSD-D), consisting of 19,166 image frames with accurately-annotated glass masks extracted from 297 videos. Extensive experiments demonstrate that VGSD-Net outperforms state-of-the-art approaches adapted from related fields. Code and dataset will be available at https://github.com/fawnliu/VGSD.

Diffusion models trained on large-scale datasets have achieved remarkable progress in image synthesis. How-ever, due to the randomness in the diffusion process, they often struggle with handling diverse low-level tasks that require details preservation. To overcome this limitation, we present a new Diff-Plugin framework to enable a single pre-trained diffusion model to generate high-fidelity re-sults across a variety of low-level tasks. Specifically, we first propose a lightweight Task-Plugin module with a dual branch design to provide task-specific priors, guiding the diffusion process in preserving image content. We then propose a Plugin-Selector that can automatically select different Task-Plugins based on the text instruction, allowing users to edit images by indicating multiple low-level tasks with natural language. We conduct extensive experiments on 8 low-level vision tasks. The results demonstrate the superiority of Diff-Plugin over existing methods, particu-larly in real-world scenarios. Our ablations further validate that Diff-Plugin is stable, schedulable, and supports robust training across different dataset sizes. Project page: https://yuhaoliu7456.github.ioIDiff-Plugin

This paper explores the weakly-supervised referring image segmentation (WRIS) problem, and focuses on a challenging setup where target localization is learned directly from image-text pairs. We note that the input text description typically already contains detailed information on how to localize the target object, and we also observe that humans often follow a step-by-step comprehension process (\ie, progressively utilizing target-related attributes and relations as cues) to identify the target object. Hence, we propose a novel Progressive Comprehension Network (PCNet) to leverage target-related textual cues from the input description for progressively localizing the target object. Specifically, we first use a Large Language Model (LLM) to decompose the input text description into short phrases. These short phrases are taken as target-related cues and fed into a Conditional Referring Module (CRM) in multiple stages, to allow updating the referring text embedding and enhance the response map for target localization in a multi-stage manner. Based on the CRM, we then propose a Region-aware Shrinking (RaS) loss to constrain the visual localization to be conducted progressively in a coarse-to-fine manner across different stages. Finally, we introduce an Instance-aware Disambiguation (IaD) loss to suppress instance localization ambiguity by differentiating overlapping response maps generated by different referring texts on the same image. Extensive experiments show that our method outperforms SOTA methods on three common benchmarks.

Quantitative infarct estimation is crucial for diagnosis, treatment and prognosis in acute ischemic stroke (AIS) patients. As the early changes of ischemic tissue are subtle and easily confounded by normal brain tissue, it remains a very challenging task. However, existing methods often ignore or confuse the contribution of different types of anatomical asymmetry caused by intrinsic and pathological changes to segmentation. Further, inefficient domain knowledge utilization leads to mis-segmentation for AIS infarcts. Inspired by this idea, we propose a pathological asymmetry-guided progressive learning (PAPL) method for AIS infarct segmentation. PAPL mimics the step-by-step learning patterns observed in humans, including three progressive stages: knowledge preparation stage, formal learning stage, and examination improvement stage. First, knowledge preparation stage accumulates the preparatory domain knowledge of the infarct segmentation task, helping to learn domain-specific knowledge representations to enhance the discriminative ability for pathological asymmetries by constructed contrastive learning task. Then, formal learning stage efficiently performs end-to-end training guided by learned knowledge representations, in which the designed feature compensation module (FCM) can leverage the anatomy similarity between adjacent slices from the volumetric medical image to help aggregate rich anatomical context information. Finally, examination improvement stage encourages improving the infarct prediction from the previous stage, where the proposed perception refinement strategy (RPRS) further exploits the bilateral difference comparison to correct the mis-segmentation infarct regions by adaptively regional shrink and expansion. Extensive experiments on public and in-house NCCT datasets demonstrated the superiority of the proposed PAPL, which is promising to help better stroke evaluation and treatment.

In this paper, we present a Neural Preset technique to address the limitations of existing color style transfer methods, including visual artifacts, vast memory requirement, and slow style switching speed. Our method is based on two core designs. First, we propose Deterministic Neural Color Mapping (DNCM) to consistently operate on each pixel via an image-adaptive color mapping matrix, avoiding artifacts and supporting high-resolution inputs with a small memory footprint. Second, we develop a two-stage pipeline by dividing the task into color normalization and stylization, which allows efficient style switching by extracting color styles as presets and reusing them on normalized input images. Due to the unavailability of pairwise datasets, we describe how to train Neural Preset via a self-supervised strategy. Various advantages of Neural Preset over existing methods are demonstrated through comprehensive evaluations. Besides, we show that our trained model can naturally support multiple applications without fine-tuning, including low-light image enhancement, underwater image correction, image dehazing, and image harmonization. The project page is: https://ZHKKKe.github.io/NeuralPreset.

Existing Referring Image Segmentation (RIS) methods typically require expensive pixel-level or box-level annotations for supervision. In this paper, we observe that the referring texts used in RIS already provide sufficient information to localize the target object. Hence, we propose a novel weakly-supervised RIS framework to formulate the target localization problem as a classification process to differentiate between positive and negative text expressions. While the referring text expressions for an image are used as positive expressions, the referring text expressions from other images can be used as negative expressions for this image. Our framework has three main novelties. First, we propose a bilateral prompt method to facilitate the classification process, by harmonizing the domain discrepancy between visual and linguistic features. Second, we propose a calibration method to reduce noisy background information and improve the correctness of the response maps for target object localization. Third, we propose a positive response map selection strategy to generate high-quality pseudo-labels from the enhanced response maps, for training a segmentation network for RIS inference. For evaluation, we propose a new metric to measure localization accuracy. Experiments on four benchmarks show that our framework achieves promising performances to existing fully-supervised RIS methods while outperforming state-of-the-art weakly-supervised methods adapted from related areas. Code is available at https://github.com/fawnliu/TRIS.

Although deep learning-based (DL-based) image processing algorithms have achieved superior performance, they are still difficult to apply on mobile devices (e.g., smartphones and cameras) due to the following reasons: 1) the high memory demand and 2) large model size. To adapt DL-based methods to mobile devices, motivated by the characteristics of image signal processors (ISPs), we propose a novel algorithm named LineDL. In LineDL, the default mode of the whole-image processing is reformulated as a line-by-line mode, eliminating the need to store large amounts of intermediate data for the whole image. An information transmission module (ITM) is designed to extract and convey the interline correlation and integrate the interline features. Furthermore, we develop a model compression method to reduce the model size while maintaining competitive performance; that is, knowledge is redefined, and compression is performed in two directions. We evaluate LineDL on general image processing tasks, including denoising and superresolution. The extensive experimental results demonstrate that LineDL achieves image quality comparable to that of state-of-the-art (SOTA) DL-based algorithms with a much smaller memory demand and competitive model size.

Existing deep learning-based shadow removal methods still produce images with shadow remnants. These shadow remnants typically exist in homogeneous regions with low-intensity values, making them untraceable in the existing image-to-image mapping paradigm. We observe that shadows mainly degrade images at the image-structure level (in which humans perceive object shapes and continuous colors). Hence, in this paper, we propose to remove shadows at the image structure level. Based on this idea, we propose a novel structure-informed shadow removal network (StructNet) to leverage the image-structure information to address the shadow remnant problem. Specifically, StructNet first reconstructs the structure information of the input image without shadows and then uses the restored shadow-free structure prior to guiding the image-level shadow removal. StructNet contains two main novel modules: 1) a mask-guided shadow-free extraction (MSFE) module to extract image structural features in a non-shadow-to-shadow directional manner; and 2) a multi-scale feature & residual aggregation (MFRA) module to leverage the shadow-free structure information to regularize feature consistency. In addition, we also propose to extend StructNet to exploit multi-level structure information (MStructNet), to further boost the shadow removal performance with minimum computational overheads. Extensive experiments on three shadow removal benchmarks demonstrate that our method outperforms existing shadow removal methods, and our StructNet can be integrated with existing methods to improve them further.

Surprising performance has been achieved in style transfer since deep learning was introduced to it. However, the existing state-of-the-art (SOTA) algorithms either suffer from quality issues or high computational complexity. The quality issues include shape retention and the adequacy of style migration, and the computational complexity is reflected in the network complexity and additional updates when the style changes. To deal with the above problems, we propose a novel low computational complexity arbitrary style transfer algorithm (LCCStyle) that mainly consists of a transformation feature module (TFM) and learning transformation module (LTM). The TFM is responsible for transforming the content feature map into the stylized feature map without impact on the integrity of content information, which contributes to good shape retention and full style migration. In addition, to avoid additional updates when the style changes, we propose a new training mechanism for arbitrary style transfer to directly generate the parameters of the TFM by a hyper-network. However, the widely used hyper-networks are composed of fully connected layers, which cause a large number of parameters. Hence, we designed a hyper-network (LTM) consisting of one-dimensional convolution to adapt to the characteristics of the Gram matrix of the style feature map, contributing to a small model size and having no impact on quality. Quantitative comparison and user study show that LCCStyle achieves high performance both on the adequacy of style migration and shape retention. Furthermore, compared with the SOTAs, the size of the proposed model is reduced by a large margin of nearly 51.4%$\sim$99.6%. When the input is 512×512 pixels, the processing speeds in the cases of unchanged style and constantly changing style are increased by at least 135% and 227%, respectively. On an Nvidia TITAN RTX GPU, LCCStyle reaches 60fps for 720p video and takes only 1 s to process 8 K images.https://github.com/HuangYujie94/LCCStyle.

Most matting research resorts to advanced semantics to achieve high-quality alpha mattes, and a direct low-level features combination is usually explored to complement alpha details. However, we argue that appearance-agnostic integration can only provide biased foreground (FG) details and that alpha mattes require different-level feature aggregation for better pixel-wise opacity perception. In this article, we propose an end-to-end hierarchical and progressive attention matting network (HAttMatting++), which can better predict the opacity of the FG from single RGB images without additional input. Specifically, we utilize channel-wise attention (CA) to distill pyramidal features and employ spatial attention (SA) at different levels to filter appearance cues. This progressive attention mechanism can estimate alpha mattes from adaptive semantics and semantics-indicated boundaries. We also introduce a hybrid loss function fusing structural similarity, mean square error, adversarial loss, and sentry supervision to guide the network to further improve the overall FG structure. In addition, we construct a large-scale and challenging image matting dataset comprised of 59,000 training images and 1,000 test images (a total of 646 distinct FG alpha mattes), which can further improve the robustness of our hierarchical and progressive aggregation model. Extensive experiments demonstrate that the proposed HAttMatting++ can capture sophisticated FG structures and achieve state-of-the-art performance with single RGB images as input.

Recently, Deep Convolutional Neural Networks (DCNNs) have achieved remarkable progress in computer vision community, including in style transfer tasks. Normally, most methods feed the full image to the DCNN. Although high-quality results can be achieved in this manner, several underlying problems arise. For one, with the increase in image resolution, the memory footprint will increase dramatically, leading to high latency and massive power consumption. Furthermore, these methods are usually unable to integrate with the commercial image signal processor (ISP), which processes the image in a line-sequential manner. To solve the above problems, we propose a novel ISP-friendly deep learning-based style transfer algorithm: SequentialStyle. A brand new line-sequential processing mode is proposed, where the image is torn into strips, and each strip is sequentially processed, contributing to less memory demand. We further propose a Spatial-Temporal Synergistic (STS) mechanism that decouples the previously simplex 2-D image style transfer into spatial feature processing (in-strip) and temporal correlation transmission (in-between strips). Compared with the SOTA style transfer algorithms, experimental results show that our SequentialStyle is competitive. Besides, SequentialStyle has less demand for memory consumption, even for the images whose resolutions are 4 k or higher.

Image matting is an ill-posed problem that aims to estimate the opacity of foreground pixels in an image. However, most existing deep learning-based methods still suffer from the coarse-grained details. In general, these algorithms are incapable of felicitously distinguishing the degree of exploration between deterministic domains (e.g.certain FG and BG pixels) and undetermined domains (e.g.uncertain in-between pixels), or inevitably lose information in the continuous sampling process, leading to a sub-optimal result. In this paper, we propose a novel network named Prior-Induced Information Alignment Matting Network (PIIAMatting), which can efficiently model the distinction of pixel-wise response maps and the correlation of layer-wise feature maps. It mainly consists of a Dynamic Gaussian Modulation mechanism (DGM) and an Information Alignment strategy (IA). Specifically, the DGM can dynamically acquire a pixel-wise domain response map learned from the prior distribution. The response map can present the relationship between the opacity variation and the convergence process during training. On the other hand, the IA comprises an Information Match Module (IMM) and an Information Aggregation Module (IAM), jointly scheduled to match and aggregate the adjacent layer-wise features adaptively. Besides, we also develop a Multi-Scale Refinement (MSR) module to integrate multi-scale receptive field information at the refinement stage to recover the fluctuating appearance details. Extensive quantitative and qualitative evaluations demonstrate that the proposed PIIAMatting performs favourably against state-of-the-art image matting methods on theAlphamatting.com, Composition-1 K and Distinctions-646 dataset.

With the development of deep convolutional neural networks, image matting has ushered in a new phase. Regarding the nature of image matting, most researches have focused on solutions for transition regions. However, we argue that many existing approaches are excessively focused on transition-dominant local fields and ignored the inherent coordination between global information and transition optimisation. In this paper, we propose the Tripartite Information Mining and Integration Network (TIMI-Net) to harmonize the coordination between global and local attributes formally. Specifically, we resort to a novel 3-branch encoder to accomplish comprehensive mining of the input information, which can supplement the neglected coordination between global and local fields. In order to achieve effective and complete interaction between such multi-branches information, we develop the Tripartite Information Integration (T I2) Module to transform and integrate the interconnections between the different branches. In addition, we built a large-scale human matting dataset (Human-2K) to advance human image matting, which consists of 2100 high-precision human images (2000 images for training and 100 images for test). Finally, we con-duct extensive experiments to prove the performance of our proposed TIMI-Net, which demonstrates that our method performs favourably against the SOTA approaches on the alphamatting.com (Rank First), Composition-1K (MSE- 0.006, Grad-11.5), Distinctions-646 and our Human-2K. Also, we have developed an online evaluation website to perform natural image matting.

Existing deep learning based matting algorithms primarily resort to high-level semantic features to improve the overall structure of alpha mattes. However, we argue that advanced semantics extracted from CNNs contribute unequally for alpha perception and we are supposed to reconcile advanced semantic information with low-level appearance cues to refine the foreground details. In this paper, we propose an end-to-end Hierarchical Attention Matting Network (HAttMatting), which can predict the better structure of alpha mattes from single RGB images without additional input. Specifically, we employ spatial and channel-wise attention to integrate appearance cues and pyramidal features in a novel fashion. This blended attention mechanism can perceive alpha mattes from refined boundaries and adaptive semantics. We also introduce a hybrid loss function fusing Structural SIMilarity (SSIM), Mean Square Error (MSE) and Adversarial loss to guide the network to further improve the overall foreground structure. Besides, we construct a large-scale image matting dataset comprised of 59,600 training images and 1000 test images (total 646 distinct foreground alpha mattes), which can further improve the robustness of our hierarchical structure aggregation model. Extensive experiments demonstrate that the proposed HAttMatting can capture sophisticated foreground structure and achieve state-of-the-art performance with single RGB images as input.

Abstract Image matting is a long‐standing problem in computer graphics and vision, mostly identified as the accurate estimation of the foreground in input images. We argue that the foreground objects can be represented by different‐level information, including the central bodies, large‐grained boundaries, refined details, etc. Based on this observation, in this paper, we propose a multi‐scale information assembly framework (MSIA‐matte) to pull out high‐quality alpha mattes from single RGB images. Technically speaking, given an input image, we extract advanced semantics as our subject content and retain initial CNN features to encode different‐level foreground expression, then combine them by our well‐designed information assembly strategy. Extensive experiments can prove the effectiveness of the proposed MSIA‐matte, and we can achieve state‐of‐the‐art performance compared to most existing matting networks.