In this paper, we present a light-weight detection transformer, LW-DETR, which outperforms YOLOs for real-time object detection. The architecture is a simple stack of a ViT encoder, a projector, and a shallow DETR decoder. Our approach leverages recent advanced techniques, such as training-effective techniques, e.g., improved loss and pretraining, and interleaved window and global attentions for reducing the ViT encoder complexity. We improve the ViT encoder by aggregating multi-level feature maps, and the intermediate and final feature maps in the ViT encoder, forming richer feature maps, and introduce window-major feature map organization for improving the efficiency of interleaved attention computation. Experimental results demonstrate that the proposed approach is superior over existing real-time detectors, e.g., YOLO and its variants, on COCO and other benchmark datasets. Code and models are available at (https://github.com/Atten4Vis/LW-DETR).
This paper introduces OpenGaussian, a method based on 3D Gaussian Splatting (3DGS) capable of 3D point-level open vocabulary understanding. Our primary motivation stems from observing that existing 3DGS-based open vocabulary methods mainly focus on 2D pixel-level parsing. These methods struggle with 3D point-level tasks due to weak feature expressiveness and inaccurate 2D-3D feature associations. To ensure robust feature presentation and 3D point-level understanding, we first employ SAM masks without cross-frame associations to train instance features with 3D consistency. These features exhibit both intra-object consistency and inter-object distinction. Then, we propose a two-stage codebook to discretize these features from coarse to fine levels. At the coarse level, we consider the positional information of 3D points to achieve location-based clustering, which is then refined at the fine level. Finally, we introduce an instance-level 3D-2D feature association method that links 3D points to 2D masks, which are further associated with 2D CLIP features. Extensive experiments, including open vocabulary-based 3D object selection, 3D point cloud understanding, click-based 3D object selection, and ablation studies, demonstrate the effectiveness of our proposed method. Project page: https://3d-aigc.github.io/OpenGaussian
Text-rich images have significant and extensive value, deeply integrated into various aspects of human life. Notably, both visual cues and linguistic symbols in text-rich images play crucial roles in information transmission but are accompanied by diverse challenges. Therefore, the efficient and effective understanding of text-rich images is a crucial litmus test for the capability of Vision-Language Models. We have crafted an efficient vision-language model, StrucTexTv3, tailored to tackle various intelligent tasks for text-rich images. The significant design of StrucTexTv3 is presented in the following aspects: Firstly, we adopt a combination of an effective multi-scale reduced visual transformer and a multi-granularity token sampler (MG-Sampler) as a visual token generator, successfully solving the challenges of high-resolution input and complex representation learning for text-rich images. Secondly, we enhance the perception and comprehension abilities of StrucTexTv3 through instruction learning, seamlessly integrating various text-oriented tasks into a unified framework. Thirdly, we have curated a comprehensive collection of high-quality text-rich images, abbreviated as TIM-30M, encompassing diverse scenarios like incidental scenes, office documents, web pages, and screenshots, thereby improving the robustness of our model. Our method achieved SOTA results in text-rich image perception tasks, and significantly improved performance in comprehension tasks. Among multimodal models with LLM decoder of approximately 1.8B parameters, it stands out as a leader, which also makes the deployment of edge devices feasible. In summary, the StrucTexTv3 model, featuring efficient structural design, outstanding performance, and broad adaptability, offers robust support for diverse intelligent application tasks involving text-rich images, thus exhibiting immense potential for widespread application.
Existing OCR engines or document image analysis systems typically rely on training separate models for text detection in varying scenarios and granularities, leading to significant computational complexity and resource demands. In this paper, we introduce "Detect Any Text" (DAT), an advanced paradigm that seamlessly unifies scene text detection, layout analysis, and document page detection into a cohesive, end-to-end model. This design enables DAT to efficiently manage text instances at different granularities, including *word*, *line*, *paragraph* and *page*. A pivotal innovation in DAT is the across-granularity interactive attention module, which significantly enhances the representation learning of text instances at varying granularities by correlating structural information across different text queries. As a result, it enables the model to achieve mutually beneficial detection performances across multiple text granularities. Additionally, a prompt-based segmentation module refines detection outcomes for texts of arbitrary curvature and complex layouts, thereby improving DAT's accuracy and expanding its real-world applicability. Experimental results demonstrate that DAT achieves state-of-the-art performances across a variety of text-related benchmarks, including multi-oriented/arbitrarily-shaped scene text detection, document layout analysis and page detection tasks.
The ultimate goal of generative models is to characterize the data distribution perfectly. For image generation, common metrics of visual quality (e.g., FID), and the truthlikeness of generated images to the human eyes seem to suggest that we are close to achieving it. However, through distribution classification tasks, we find that, in the eyes of classifiers parameterized by neural networks, the strongest diffusion models are still far from this goal. Specifically, classifiers consistently and effortlessly distinguish between real and generated images in various settings. Further, we observe an intriguing discrepancy: classifiers can identify differences between diffusion models with similar performance (e.g., U-ViT-H vs. DiT-XL), but struggle to differentiate between the smallest and largest models in the same family (e.g., EDM2-XS vs. EDM2-XXL), whereas humans exhibit the opposite tendency. As an explanation, our comprehensive empirical study suggests that, unlike humans, classifiers tend to classify images through edge and high-frequency components. We believe that our methodology can serve as a probe to understand how generative models work and inspire further thought on how existing models can be improved and how the abuse of such models can be prevented.
In recent times, following the paradigm of DETR (DEtection TRansformer), query-based end-to-end instance segmentation (QEIS) methods have exhibited superior performance compared to CNN-based models, particularly when trained on large-scale datasets. Nevertheless, the effectiveness of these QEIS methods diminishes significantly when confronted with limited training data. This limitation arises from their reliance on substantial data volumes to effectively train the pivotal queries/kernels that are essential for acquiring localization and shape priors. To address this problem, we propose a novel method for unsupervised pre-training in low-data regimes. Inspired by the recently successful prompting technique, we introduce a new method, Unsupervised Pre-training with Language-Vision Prompts (UPLVP), which improves QEIS models' instance segmentation by bringing language-vision prompts to queries/kernels. Our method consists of three parts: (1) Masks Proposal: Utilizes language-vision models to generate pseudo masks based on unlabeled images. (2) Prompt-Kernel Matching: Converts pseudo masks into prompts and injects the best-matched localization and shape features to their corresponding kernels. (3) Kernel Supervision: Formulates supervision for pre-training at the kernel level to ensure robust learning. With the help of our pre-training method, QEIS models can converge faster and perform better than CNN-based models in low-data regimes. Experimental evaluations conducted on MS COCO, Cityscapes, and CTW1500 datasets indicate that the QEIS models' performance can be significantly improved when pre-trained with our method. Code will be available at: https://github.com/lifuguan/UPLVP.
Do we fully leverage the potential of visual encoder in Multimodal Large Language Models (MLLMs)? The recent outstanding performance of MLLMs in multimodal understanding has garnered broad attention from both academia and industry. In the current MLLM rat race, the focus seems to be predominantly on the linguistic side. We witness the rise of larger and higher-quality instruction datasets, as well as the involvement of larger-sized LLMs. Yet, scant attention has been directed towards the visual signals utilized by MLLMs, often assumed to be the final high-level features extracted by a frozen visual encoder. In this paper, we introduce the Dense Connector - a simple, effective, and plug-and-play vision-language connector that significantly enhances existing MLLMs by leveraging multi-layer visual features, with minimal additional computational overhead. Furthermore, our model, trained solely on images, showcases remarkable zero-shot capabilities in video understanding as well. Experimental results across various vision encoders, image resolutions, training dataset scales, varying sizes of LLMs (2.7B->70B), and diverse architectures of MLLMs (e.g., LLaVA and Mini-Gemini) validate the versatility and scalability of our approach, achieving state-of-the-art performance on across 19 image and video benchmarks. We hope that this work will provide valuable experience and serve as a basic module for future MLLM development.
We present Real-time Gaussian SLAM (RTG-SLAM), a real-time 3D reconstruction system with an RGBD camera for large-scale environments using Gaussian splatting. The system features a compact Gaussian representation and a highly efficient on-the-fly Gaussian optimization scheme. We force each Gaussian to be either opaque or nearly transparent, with the opaque ones fitting the surface and dominant colors, and transparent ones fitting residual colors. By rendering depth in a different way from color rendering, we let a single opaque Gaussian well fit a local surface region without the need of multiple overlapping Gaussians, hence largely reducing the memory and computation cost. For on-the-fly Gaussian optimization, we explicitly add Gaussians for three types of pixels per frame: newly observed, with large color errors, and with large depth errors. We also categorize all Gaussians into stable and unstable ones, where the stable Gaussians are expected to well fit previously observed RGBD images and otherwise unstable. We only optimize the unstable Gaussians and only render the pixels occupied by unstable Gaussians. In this way, both the number of Gaussians to be optimized and pixels to be rendered are largely reduced, and the optimization can be done in real time. We show real-time reconstructions of a variety of large scenes. Compared with the state-of-the-art NeRF-based RGBD SLAM, our system achieves comparable high-quality reconstruction but with around twice the speed and half the memory cost, and shows superior performance in the realism of novel view synthesis and camera tracking accuracy.
Prompt learning has become the most effective paradigm for adapting large pre-trained vision-language models (VLMs) to downstream tasks. Recently, unsupervised prompt tuning methods, such as UPL and POUF, directly leverage pseudo-labels as supervisory information to fine-tune additional adaptation modules on unlabeled data. However, inaccurate pseudo labels easily misguide the tuning process and result in poor representation capabilities. In light of this, we propose Training-Free Unsupervised Prompts (TFUP), which maximally preserves the inherent representation capabilities and enhances them with a residual connection to similarity-based prediction probabilities in a training-free and labeling-free manner. Specifically, we integrate both instance confidence and prototype scores to select representative samples, which are used to customize a reliable Feature Cache Model (FCM) for training-free inference. Then, we design a Multi-level Similarity Measure (MSM) that considers both feature-level and semantic-level similarities to calculate the distance between each test image and the cached sample as the weight of the corresponding cached label to generate similarity-based prediction probabilities. In this way, TFUP achieves surprising performance, even surpassing the training-base method on multiple classification datasets. Based on our TFUP, we propose a training-based approach (TFUP-T) to further boost the adaptation performance. In addition to the standard cross-entropy loss, TFUP-T adopts an additional marginal distribution entropy loss to constrain the model from a global perspective. Our TFUP-T achieves new state-of-the-art classification performance compared to unsupervised and few-shot adaptation approaches on multiple benchmarks. In particular, TFUP-T improves the classification accuracy of POUF by 3.3% on the most challenging Domain-Net dataset.
The recent 3D Gaussian Splatting (GS) exhibits high-quality and real-time synthesis of novel views in 3D scenes. Currently, it primarily focuses on geometry and appearance modeling, while lacking the semantic understanding of scenes. To bridge this gap, we present CLIP-GS, which integrates semantics from Contrastive Language-Image Pre-Training (CLIP) into Gaussian Splatting to efficiently comprehend 3D environments without annotated semantic data. In specific, rather than straightforwardly learning and rendering high-dimensional semantic features of 3D Gaussians, which significantly diminishes the efficiency, we propose a Semantic Attribute Compactness (SAC) approach. SAC exploits the inherent unified semantics within objects to learn compact yet effective semantic representations of 3D Gaussians, enabling highly efficient rendering (>100 FPS). Additionally, to address the semantic ambiguity, caused by utilizing view-inconsistent 2D CLIP semantics to supervise Gaussians, we introduce a 3D Coherent Self-training (3DCS) strategy, resorting to the multi-view consistency originated from the 3D model. 3DCS imposes cross-view semantic consistency constraints by leveraging refined, self-predicted pseudo-labels derived from the trained 3D Gaussian model, thereby enhancing precise and view-consistent segmentation results. Extensive experiments demonstrate that our method remarkably outperforms existing state-of-the-art approaches, achieving improvements of 17.29% and 20.81% in mIoU metric on Replica and ScanNet datasets, respectively, while maintaining real-time rendering speed. Furthermore, our approach exhibits superior performance even with sparse input data, verifying the robustness of our method.