Lixiang Ru

Lixiang Ru, Yibing Zhan, Baosheng Yu et al.

Weakly-supervised semantic segmentation (WSSS) with image-level labels is an important and challenging task. Due to the high training efficiency, end-to-end solutions for WSSS have received increasing attention from the community. However, current methods are mainly based on convolutional neural networks and fail to explore the global information properly, thus usually resulting in incomplete object regions. In this paper, to address the aforementioned problem, we introduce Transformers, which naturally integrate global information, to generate more integral initial pseudo labels for end-to-end WSSS. Motivated by the inherent consistency between the self-attention in Transformers and the semantic affinity, we propose an Affinity from Attention (AFA) module to learn semantic affinity from the multi-head self-attention (MHSA) in Transformers. The learned affinity is then leveraged to refine the initial pseudo labels for segmentation. In addition, to efficiently derive reliable affinity labels for supervising AFA and ensure the local consistency of pseudo labels, we devise a Pixel-Adaptive Refinement module that incorporates low-level image appearance information to refine the pseudo labels. We perform extensive experiments and our method achieves 66.0% and 38.9% mIoU on the PASCAL VOC 2012 and MS COCO 2014 datasets, respectively, significantly outperforming recent end-to-end methods and several multi-stage competitors. Code is available at https://github.com/rulixiang/afa.

Lixiang Ru, Heliang Zheng, Yibing Zhan et al.

Weakly-Supervised Semantic Segmentation (WSSS) using image-level labels typically utilizes Class Activation Map (CAM) to generate the pseudo labels. Limited by the local structure perception of CNN, CAM usually cannot identify the integral object regions. Though the recent Vision Transformer (ViT) can remedy this flaw, we observe it also brings the over-smoothing issue, \ie, the final patch tokens incline to be uniform. In this work, we propose Token Contrast (ToCo) to address this issue and further explore the virtue of ViT for WSSS. Firstly, motivated by the observation that intermediate layers in ViT can still retain semantic diversity, we designed a Patch Token Contrast module (PTC). PTC supervises the final patch tokens with the pseudo token relations derived from intermediate layers, allowing them to align the semantic regions and thus yield more accurate CAM. Secondly, to further differentiate the low-confidence regions in CAM, we devised a Class Token Contrast module (CTC) inspired by the fact that class tokens in ViT can capture high-level semantics. CTC facilitates the representation consistency between uncertain local regions and global objects by contrasting their class tokens. Experiments on the PASCAL VOC and MS COCO datasets show the proposed ToCo can remarkably surpass other single-stage competitors and achieve comparable performance with state-of-the-art multi-stage methods. Code is available at https://github.com/rulixiang/ToCo.

Zhenghui Zhao, Lixiang Ru, Chen Wu

Weakly-supervised change detection (WSCD) aims to detect pixel-level changes with only image-level annotations. Owing to its label efficiency, WSCD is drawing increasing attention recently. However, current WSCD methods often encounter the challenge of change missing and fabricating, i.e., the inconsistency between image-level annotations and pixel-level predictions. Specifically, change missing refer to the situation that the WSCD model fails to predict any changed pixels, even though the image-level label indicates changed, and vice versa for change fabricating. To address this challenge, in this work, we leverage global-scale and local-scale priors in WSCD and propose two components: a Dilated Prior (DP) decoder and a Label Gated (LG) constraint. The DP decoder decodes samples with the changed image-level label, skips samples with the unchanged label, and replaces them with an all-unchanged pixel-level label. The LG constraint is derived from the correspondence between changed representations and image-level labels, penalizing the model when it mispredicts the change status. Additionally, we develop TransWCD, a simple yet powerful transformer-based model, showcasing the potential of weakly-supervised learning in change detection. By integrating the DP decoder and LG constraint into TransWCD, we form TransWCD-DL. Our proposed TransWCD and TransWCD-DL achieve significant +6.33% and +9.55% F1 score improvements over the state-of-the-art methods on the WHU-CD dataset, respectively. Some performance metrics even exceed several fully-supervised change detection (FSCD) competitors. Code will be available at https://github.com/zhenghuizhao/TransWCD.

Yingying Zhang, Xin Guo, Jiangwei Lao et al.

Large-scale self-supervised pre-training has paved the way for one foundation model to handle many different vision tasks. Most pre-training methodologies train a single model of a certain size at one time. Nevertheless, various computation or storage constraints in real-world scenarios require substantial efforts to develop a series of models with different sizes to deploy. Thus, in this study, we propose a novel tri-branch self-supervised training framework, termed as POA (Pre-training Once for All), to tackle this aforementioned issue. Our approach introduces an innovative elastic student branch into a modern self-distillation paradigm. At each pre-training step, we randomly sample a sub-network from the original student to form the elastic student and train all branches in a self-distilling fashion. Once pre-trained, POA allows the extraction of pre-trained models of diverse sizes for downstream tasks. Remarkably, the elastic student facilitates the simultaneous pre-training of multiple models with different sizes, which also acts as an additional ensemble of models of various sizes to enhance representation learning. Extensive experiments, including k-nearest neighbors, linear probing evaluation and assessments on multiple downstream tasks demonstrate the effectiveness and advantages of our POA. It achieves state-of-the-art performance using ViT, Swin Transformer and ResNet backbones, producing around a hundred models with different sizes through a single pre-training session. The code is available at: https://github.com/Qichuzyy/POA.

Inclusion AI, Biao Gong, Cheng Zou et al.

We propose Ming-Omni, a unified multimodal model capable of processing images, text, audio, and video, while demonstrating strong proficiency in both speech and image generation. Ming-Omni employs dedicated encoders to extract tokens from different modalities, which are then processed by Ling, an MoE architecture equipped with newly proposed modality-specific routers. This design enables a single model to efficiently process and fuse multimodal inputs within a unified framework, thereby facilitating diverse tasks without requiring separate models, task-specific fine-tuning, or structural redesign. Importantly, Ming-Omni extends beyond conventional multimodal models by supporting audio and image generation. This is achieved through the integration of an advanced audio decoder for natural-sounding speech and Ming-Lite-Uni for high-quality image generation, which also allow the model to engage in context-aware chatting, perform text-to-speech conversion, and conduct versatile image editing. Our experimental results showcase Ming-Omni offers a powerful solution for unified perception and generation across all modalities. Notably, our proposed Ming-Omni is the first open-source model we are aware of to match GPT-4o in modality support, and we release all code and model weights to encourage further research and development in the community.

Inclusion AI, Biao Gong, Cheng Zou et al.

We introduce Ming-Lite-Uni, an open-source multimodal framework featuring a newly designed unified visual generator and a native multimodal autoregressive model tailored for unifying vision and language. Specifically, this project provides an open-source implementation of the integrated MetaQueries and M2-omni framework, while introducing the novel multi-scale learnable tokens and multi-scale representation alignment strategy. By leveraging a fixed MLLM and a learnable diffusion model, Ming-Lite-Uni enables native multimodal AR models to perform both text-to-image generation and instruction based image editing tasks, expanding their capabilities beyond pure visual understanding. Our experimental results demonstrate the strong performance of Ming-Lite-Uni and illustrate the impressive fluid nature of its interactive process. All code and model weights are open-sourced to foster further exploration within the community. Notably, this work aligns with concurrent multimodal AI milestones - such as ChatGPT-4o with native image generation updated in March 25, 2025 - underscoring the broader significance of unified models like Ming-Lite-Uni on the path toward AGI. Ming-Lite-Uni is in alpha stage and will soon be further refined.

Zhenghui Zhao, Chen Wu, Lixiang Ru et al.

Existing Weakly-Supervised Change Detection (WSCD) methods often encounter the problem of "instance lumping" under scene-level supervision, particularly in scenarios with a dense distribution of changed instances (i.e., changed objects). In these scenarios, unchanged pixels between changed instances are also mistakenly identified as changed, causing multiple changes to be mistakenly viewed as one. In practical applications, this issue prevents the accurate quantification of the number of changes. To address this issue, we propose a Dense Instance Separation (DISep) method as a plug-and-play solution, refining pixel features from a unified instance perspective under scene-level supervision. Specifically, our DISep comprises a three-step iterative training process: 1) Instance Localization: We locate instance candidate regions for changed pixels using high-pass class activation maps. 2) Instance Retrieval: We identify and group these changed pixels into different instance IDs through connectivity searching. Then, based on the assigned instance IDs, we extract corresponding pixel-level features on a per-instance basis. 3) Instance Separation: We introduce a separation loss to enforce intra-instance pixel consistency in the embedding space, thereby ensuring separable instance feature representations. The proposed DISep adds only minimal training cost and no inference cost. It can be seamlessly integrated to enhance existing WSCD methods. We achieve state-of-the-art performance by enhancing {three Transformer-based and four ConvNet-based methods} on the LEVIR-CD, WHU-CD, DSIFN-CD, SYSU-CD, and CDD datasets. Additionally, our DISep can be used to improve fully-supervised change detection methods. Code is available at https://github.com/zhenghuizhao/Plug-and-Play-DISep-for-Change-Detection.

Xin Guo, Jiangwei Lao, Bo Dang et al.

Prior studies on Remote Sensing Foundation Model (RSFM) reveal immense potential towards a generic model for Earth Observation. Nevertheless, these works primarily focus on a single modality without temporal and geo-context modeling, hampering their capabilities for diverse tasks. In this study, we present SkySense, a generic billion-scale model, pre-trained on a curated multi-modal Remote Sensing Imagery (RSI) dataset with 21.5 million temporal sequences. SkySense incorporates a factorized multi-modal spatiotemporal encoder taking temporal sequences of optical and Synthetic Aperture Radar (SAR) data as input. This encoder is pre-trained by our proposed Multi-Granularity Contrastive Learning to learn representations across different modal and spatial granularities. To further enhance the RSI representations by the geo-context clue, we introduce Geo-Context Prototype Learning to learn region-aware prototypes upon RSI's multi-modal spatiotemporal features. To our best knowledge, SkySense is the largest Multi-Modal RSFM to date, whose modules can be flexibly combined or used individually to accommodate various tasks. It demonstrates remarkable generalization capabilities on a thorough evaluation encompassing 16 datasets over 7 tasks, from single- to multi-modal, static to temporal, and classification to localization. SkySense surpasses 18 recent RSFMs in all test scenarios. Specifically, it outperforms the latest models such as GFM, SatLas and Scale-MAE by a large margin, i.e., 2.76%, 3.67% and 3.61% on average respectively. We will release the pre-trained weights to facilitate future research and Earth Observation applications.

Peiqi Chen, Lei Yu, Yi Wan et al.

Semi-dense feature matching methods have shown strong performance in challenging scenarios. However, the existing pipeline relies on a global search across the entire feature map to establish coarse matches, limiting further improvements in accuracy and efficiency. Motivated by this limitation, we propose a novel pipeline, CasP, which leverages cascaded correspondence priors for guidance. Specifically, the matching stage is decomposed into two progressive phases, bridged by a region-based selective cross-attention mechanism designed to enhance feature discriminability. In the second phase, one-to-one matches are determined by restricting the search range to the one-to-many prior areas identified in the first phase. Additionally, this pipeline benefits from incorporating high-level features, which helps reduce the computational costs of low-level feature extraction. The acceleration gains of CasP increase with higher resolution, and our lite model achieves a speedup of $\sim2.2\times$ at a resolution of 1152 compared to the most efficient method, ELoFTR. Furthermore, extensive experiments demonstrate its superiority in geometric estimation, particularly with impressive cross-domain generalization. These advantages highlight its potential for latency-sensitive and high-robustness applications, such as SLAM and UAV systems. Code is available at https://github.com/pq-chen/CasP.

Inclusion AI, Fudong Wang, Jiajia Liu et al.

Recent advancements in Multimodal Large Language Models (MLLMs), particularly through Reinforcement Learning with Verifiable Rewards (RLVR), have significantly enhanced their reasoning abilities. However, a critical gap persists: these models struggle with dynamic spatial interactions, a capability essential for real-world applications. To bridge this gap, we introduce M2-Reasoning-7B, a model designed to excel in both general and spatial reasoning. Our approach integrates two key innovations: (1) a novel data pipeline that generates 294.2K high-quality data samples (168K for cold-start fine-tuning and 126.2K for RLVR), which feature logically coherent reasoning trajectories and have undergone comprehensive assessment; and (2) a dynamic multi-task training strategy with step-wise optimization to mitigate conflicts between data, and task-specific rewards for delivering tailored incentive signals. This combination of curated data and advanced training allows M2-Reasoning-7B to set a new state-of-the-art (SOTA) across 8 benchmarks, showcasing superior performance in both general and spatial reasoning domains.

Yingying Zhang, Lixiang Ru, Kang Wu et al.

The multi-modal remote sensing foundation model (MM-RSFM) has significantly advanced various Earth observation tasks, such as urban planning, environmental monitoring, and natural disaster management. However, most existing approaches generally require the training of separate backbone networks for each data modality, leading to redundancy and inefficient parameter utilization. Moreover, prevalent pre-training methods typically apply self-supervised learning (SSL) techniques from natural images without adequately accommodating the characteristics of remote sensing (RS) images, such as the complicated semantic distribution within a single RS image. In this work, we present SkySense V2, a unified MM-RSFM that employs a single transformer backbone to handle multiple modalities. This backbone is pre-trained with a novel SSL strategy tailored to the distinct traits of RS data. In particular, SkySense V2 incorporates an innovative adaptive patch merging module and learnable modality prompt tokens to address challenges related to varying resolutions and limited feature diversity across modalities. In additional, we incorporate the mixture of experts (MoE) module to further enhance the performance of the foundation model. SkySense V2 demonstrates impressive generalization abilities through an extensive evaluation involving 16 datasets over 7 tasks, outperforming SkySense by an average of 1.8 points.

Xiaolong Wang, Lei Yu, Yingying Zhang et al.

Feature matching between image pairs is a fundamental problem in computer vision that drives many applications, such as SLAM. Recently, semi-dense matching approaches have achieved substantial performance enhancements and established a widely-accepted coarse-to-fine paradigm. However, the majority of existing methods focus on improving coarse feature representation rather than the fine-matching module. Prior fine-matching techniques, which rely on point-to-patch matching probability expectation or direct regression, often lack precision and do not guarantee the continuity of feature points across sequential images. To address this limitation, this paper concentrates on enhancing the fine-matching module in the semi-dense matching framework. We employ a lightweight and efficient homography estimation network to generate the perspective mapping between patches obtained from coarse matching. This patch-to-patch approach achieves the overall alignment of two patches, resulting in a higher sub-pixel accuracy by incorporating additional constraints. By leveraging the homography estimation between patches, we can achieve a dense matching result with low computational cost. Extensive experiments demonstrate that our method achieves higher accuracy compared to previous semi-dense matchers. Meanwhile, our dense matching results exhibit similar end-point-error accuracy compared to previous dense matchers while maintaining semi-dense efficiency.

Inclusion AI, Bowen Ma, Cheng Zou et al.

We propose Ming-Flash-Omni, an upgraded version of Ming-Omni, built upon a sparser Mixture-of-Experts (MoE) variant of Ling-Flash-2.0 with 100 billion total parameters, of which only 6.1 billion are active per token. This architecture enables highly efficient scaling (dramatically improving computational efficiency while significantly expanding model capacity) and empowers stronger unified multimodal intelligence across vision, speech, and language, representing a key step toward Artificial General Intelligence (AGI). Compared to its predecessor, the upgraded version exhibits substantial improvements across multimodal understanding and generation. We significantly advance speech recognition capabilities, achieving state-of-the-art performance in contextual ASR and highly competitive results in dialect-aware ASR. In image generation, Ming-Flash-Omni introduces high-fidelity text rendering and demonstrates marked gains in scene consistency and identity preservation during image editing. Furthermore, Ming-Flash-Omni introduces generative segmentation, a capability that not only achieves strong standalone segmentation performance but also enhances spatial control in image generation and improves editing consistency. Notably, Ming-Flash-Omni achieves state-of-the-art results in text-to-image generation and generative segmentation, and sets new records on all 12 contextual ASR benchmarks, all within a single unified architecture.

Panwang Xia, Lei Yu, Yi Wan et al.

Cross-View Geo-Localization tackles the challenge of image geo-localization in GNSS-denied environments, including disaster response scenarios, urban canyons, and dense forests, by matching street-view query images with geo-tagged aerial-view reference images. However, current research often relies on benchmarks and methods that assume center-aligned settings or account for only limited decentrality, which we define as the offset of the query image relative to the reference image center. Such assumptions fail to reflect real-world scenarios, where reference databases are typically pre-established without the possibility of ensuring perfect alignment for each query image. Moreover, decentrality is a critical factor warranting deeper investigation, as larger decentrality can substantially improve localization efficiency but comes at the cost of declines in localization accuracy. To address this limitation, we introduce DReSS (Decentrality Related Street-view and Satellite-view dataset), a novel dataset designed to evaluate cross-view geo-localization with a large geographic scope and diverse landscapes, emphasizing the decentrality issue. Meanwhile, we propose AuxGeo (Auxiliary Enhanced Geo-Localization) to further study the decentrality issue, which leverages a multi-metric optimization strategy with two novel modules: the Bird's-eye view Intermediary Module (BIM) and the Position Constraint Module (PCM). These modules improve the localization accuracy despite the decentrality problem. Extensive experiments demonstrate that AuxGeo outperforms previous methods on our proposed DReSS dataset, mitigating the issue of large decentrality, and also achieves state-of-the-art performance on existing public datasets such as CVUSA, CVACT, and VIGOR.

Xiaolong Wang, Lixiang Ru, Ziyuan Huang et al.

We propose a novel AutoRegressive Generation-based paradigm for image Segmentation (ARGenSeg), achieving multimodal understanding and pixel-level perception within a unified framework. Prior works integrating image segmentation into multimodal large language models (MLLMs) typically employ either boundary points representation or dedicated segmentation heads. These methods rely on discrete representations or semantic prompts fed into task-specific decoders, which limits the ability of the MLLM to capture fine-grained visual details. To address these challenges, we introduce a segmentation framework for MLLM based on image generation, which naturally produces dense masks for target objects. We leverage MLLM to output visual tokens and detokenize them into images using an universal VQ-VAE, making the segmentation fully dependent on the pixel-level understanding of the MLLM. To reduce inference latency, we employ a next-scale-prediction strategy to generate required visual tokens in parallel. Extensive experiments demonstrate that our method surpasses prior state-of-the-art approaches on multiple segmentation datasets with a remarkable boost in inference speed, while maintaining strong understanding capabilities.

Lixiang Ru, Bo Du, Yibing Zhan et al.

Weakly-Supervised Semantic Segmentation (WSSS) methods with image-level labels generally train a classification network to generate the Class Activation Maps (CAMs) as the initial coarse segmentation labels. However, current WSSS methods still perform far from satisfactorily because their adopted CAMs 1) typically focus on partial discriminative object regions and 2) usually contain useless background regions. These two problems are attributed to the sole image-level supervision and aggregation of global information when training the classification networks. In this work, we propose the visual words learning module and hybrid pooling approach, and incorporate them in the classification network to mitigate the above problems. In the visual words learning module, we counter the first problem by enforcing the classification network to learn fine-grained visual word labels so that more object extents could be discovered. Specifically, the visual words are learned with a codebook, which could be updated via two proposed strategies, i.e. learning-based strategy and memory-bank strategy. The second drawback of CAMs is alleviated with the proposed hybrid pooling, which incorporates the global average and local discriminative information to simultaneously ensure object completeness and reduce background regions. We evaluated our methods on PASCAL VOC 2012 and MS COCO 2014 datasets. Without any extra saliency prior, our method achieved 70.6% and 70.7% mIoU on the $val$ and $test$ set of PASCAL VOC dataset, respectively, and 36.2% mIoU on the $val$ set of MS COCO dataset, which significantly surpassed the performance of state-of-the-art WSSS methods.

Chen Wu, Yinong Guo, Haonan Guo et al.

In order to mitigate the spread of COVID-19, Wuhan was the first city to implement strict lockdown policy in 2020. Even though numerous researches have discussed the travel restriction between cities and provinces, few studies focus on the effect of transportation control inside the city due to the lack of the measurement and available data in Wuhan. Since the public transports have been shut down in the beginning of city lockdown, the change of traffic density is a good indicator to reflect the intracity population flow. Therefore, in this paper, we collected time-series high-resolution remote sensing images with the resolution of 1m acquired before, during and after Wuhan lockdown by GF-2 satellite. Vehicles on the road were extracted and counted for the statistics of traffic density to reflect the changes of human transmissions in the whole period of Wuhan lockdown. Open Street Map was used to obtain observation road surfaces, and a vehicle detection method combing morphology filter and deep learning was utilized to extract vehicles with the accuracy of 62.56%. According to the experimental results, the traffic density of Wuhan dropped with the percentage higher than 80%, and even higher than 90% on main roads during city lockdown; after lockdown lift, the traffic density recovered to the normal rate. Traffic density distributions also show the obvious reduction and increase throughout the whole study area. The significant reduction and recovery of traffic density indicates that the lockdown policy in Wuhan show effectiveness in controlling human transmission inside the city, and the city returned to normal after lockdown lift.

Lixiang Ru, Bo Du, Chen Wu

Classifying multi-temporal scene land-use categories and detecting their semantic scene-level changes for imagery covering urban regions could straightly reflect the land-use transitions. Existing methods for scene change detection rarely focus on the temporal correlation of bi-temporal features, and are mainly evaluated on small scale scene change detection datasets. In this work, we proposed a CorrFusion module that fuses the highly correlated components in bi-temporal feature embeddings. We firstly extracts the deep representations of the bi-temporal inputs with deep convolutional networks. Then the extracted features will be projected into a lower dimension space to computed the instance-level correlation. The cross-temporal fusion will be performed based on the computed correlation in CorrFusion module. The final scene classification are obtained with softmax activation layers. In the objective function, we introduced a new formulation for calculating the temporal correlation. The detailed derivation of backpropagation gradients for the proposed module is also given in this paper. Besides, we presented a much larger scale scene change detection dataset and conducted experiments on this dataset. The experimental results demonstrated that our proposed CorrFusion module could remarkably improve the multi-temporal scene classification and scene change detection results.

Bo Du, Lixiang Ru, Chen Wu et al.

Change detection has been a hotspot in remote sensing technology for a long time. With the increasing availability of multi-temporal remote sensing images, numerous change detection algorithms have been proposed. Among these methods, image transformation methods with feature extraction and mapping could effectively highlight the changed information and thus has better change detection performance. However, changes of multi-temporal images are usually complex, existing methods are not effective enough. In recent years, deep network has shown its brilliant performance in many fields including feature extraction and projection. Therefore, in this paper, based on deep network and slow feature analysis (SFA) theory, we proposed a new change detection algorithm for multi-temporal remotes sensing images called Deep Slow Feature Analysis (DSFA). In DSFA model, two symmetric deep networks are utilized for projecting the input data of bi-temporal imagery. Then, the SFA module is deployed to suppress the unchanged components and highlight the changed components of the transformed features. The CVA pre-detection is employed to find unchanged pixels with high confidence as training samples. Finally, the change intensity is calculated with chi-square distance and the changes are determined by threshold algorithms. The experiments are performed on two real-world datasets and a public hyperspectral dataset. The visual comparison and quantitative evaluation have both shown that DSFA could outperform the other state-of-the-art algorithms, including other SFA-based and deep learning methods.