arXiv subjects

Jiale Cao

Publications and source records attributed to Jiale Cao.

At least 19 recordsLinked to original sources

arXiv2026

UniFS: Unified Fast-to-Slow Hierarchical Architecture for Vision-Language-Action Models

Mainstream Fast-Slow dual system vision-language-action models decouple a high-frequency action expert from a low-frequency vision-language model for efficiency, yet they face a fundamental frequency dilemma: large update gaps cause semantic drift from stale context, while small gaps erode the intended computational savings. Moreover, because the action expert receives only the VLM's final-layer representation at a single fixed frequency, rich intermediate features are discarded, limiting both information coupling and manipulation precision. Inspired by multi-timescale neural processing in the human brain, we introduce UniFS, a unified fast-to-slow architecture that resolves these challenges through three key designs. First, we stratify the VLM layers into groups with progressively decreasing update frequencies, enabling shallow layers to capture fast-changing dynamics while deeper layers cache stable semantic context. Second, a latent vector inversion mechanism re-routes the interaction order between multi-scale VLM features and the action expert, aligning fast-varying representations with fine-grained action decoding and slow-varying ones with coarse planning. Third, a multi-level supervision strategy enforces a coarse-to-fine learning hierarchy across temporal scales. Together, these designs enable richer cross-frequency information transfer within a single backbone, while the low-frequency pathways additionally preserve temporal context across steps. Experiments on LIBERO show that UniFS achieves state-of-the-art performance (98.3\% average success rate, a 2.5\% gain over VLA-Adapter baseline) while reducing average inference latency from 36.5~ms to 17.8~ms (2.1$\times$ speedup). Real-robot experiments on a Franka platform further validate its practical applicability. Code is opensourced at https://github.com/linsun449/UniFS.

cs.RO↗

arXiv2026

Not All Modalities Are Equal: Instruction-Aware Gating for Multimodal Videos

Pre-trained video large language models excel at visual reasoning. However, they struggle when videos arrive with auxiliary streams, such as audio, depth map, or dense temporal evidence. In such a scenario, uniform fusion induces modality interference, allowing irrelevant channels to distract the model. To address this issue, we present a unified multimodal video understanding framework, named UniMVU, that performs instruction-aware fusion across video, audio, depth map, or any other modality inputs via two levels of dynamic gating: inner-modality gates emphasize salient regions within each modality, whereas modality-level gates re-weight whole streams; both are conditioned on the text instruction to adaptively balance modality importance. Our UniMVU combines cross-modal self-attention with instruction-driven inner-modality gating module and a modality-level gating module with control token; for time-aligned streams we further adopt a fast-to-slow fusion scheme that reduces redundancy. Across six benchmarks (AVQA, AVSD, Music-AVQA, ScanQA, SQA3D and MVBench), our UniMVU achieves consistent gains over static-fusion baselines achieving gains as high as 13.5 in terms of CIDEr metric. Further, our analysis shows that the gating mechanism aligns with the human-interpretable modality relevance, and ablations show the contributions of inner-modality and modality-level gating. Our UniMVU provides a simple, unified recipe for instruction-aware multimodal video understanding that scales to diverse modalities without hand-crafted fusion rules.

cs.CV↗

arXiv2026

STDec: Spatio-Temporal Stability Guided Decoding for dLLMs

Diffusion Large Language Models (dLLMs) have achieved rapid progress, viewed as a promising alternative to the autoregressive paradigm. However, most dLLM decoders still adopt a global confidence threshold, and do not explicitly model local context from neighboring decoded states or temporal consistency of predicted token IDs across steps. To address this issue, we propose a simple spatio-temporal stability guided decoding approach, named STDec. We observe strong spatio-temporal stability in dLLM decoding: newly decoded tokens tend to lie near decoded neighbors, and their predicted IDs often remain consistent across several denoising steps. Inspired by this stability, our STDec includes spatial-aware decoding and temporal-aware decoding. The spatial-aware decoding dynamically generates the token-adaptive threshold by aggregating the decoded states of nearby tokens. The temporal-aware decoding relaxes the decoding thresholds for tokens whose predicted token IDs remain consistent over denoising steps. Our STDec is training-free and remains compatible with cache-based acceleration methods. Across textual reasoning and multimodal understanding benchmarks, STDec substantially improves throughput while maintaining comparable task performance score. Notably, on MBPP with LLaDA, STDec achieves up to 14.17x speedup with a comparable score. Homepage: https://yzchen02.github.io/STDec.

cs.CL↗

arXiv2025

SNNSIR: A Simple Spiking Neural Network for Stereo Image Restoration

Spiking Neural Networks (SNNs), characterized by discrete binary activations, offer high computational efficiency and low energy consumption, making them well-suited for computation-intensive tasks such as stereo image restoration. In this work, we propose SNNSIR, a simple yet effective Spiking Neural Network for Stereo Image Restoration, specifically designed under the spike-driven paradigm where neurons transmit information through sparse, event-based binary spikes. In contrast to existing hybrid SNN-ANN models that still rely on operations such as floating-point matrix division or exponentiation, which are incompatible with the binary and event-driven nature of SNNs, our proposed SNNSIR adopts a fully spike-driven architecture to achieve low-power and hardware-friendly computation. To address the expressiveness limitations of binary spiking neurons, we first introduce a lightweight Spike Residual Basic Block (SRBB) to enhance information flow via spike-compatible residual learning. Building on this, the Spike Stereo Convolutional Modulation (SSCM) module introduces simplified nonlinearity through element-wise multiplication and highlights noise-sensitive regions via cross-view-aware modulation. Complementing this, the Spike Stereo Cross-Attention (SSCA) module further improves stereo correspondence by enabling efficient bidirectional feature interaction across views within a spike-compatible framework. Extensive experiments on diverse stereo image restoration tasks, including rain streak removal, raindrop removal, low-light enhancement, and super-resolution demonstrate that our model achieves competitive restoration performance while significantly reducing computational overhead. These results highlight the potential for real-time, low-power stereo vision applications. The code will be available after the article is accepted.

cs.CV↗

arXiv2025

GeoVLA: Empowering 3D Representations in Vision-Language-Action Models

Vision-Language-Action (VLA) models have emerged as a promising approach for enabling robots to follow language instructions and predict corresponding actions. However, current VLA models mainly rely on 2D visual inputs, neglecting the rich geometric information in the 3D physical world, which limits their spatial awareness and adaptability. In this paper, we present GeoVLA, a novel VLA framework that effectively integrates 3D information to advance robotic manipulation. It uses a vision-language model (VLM) to process images and language instructions,extracting fused vision-language embeddings. In parallel, it converts depth maps into point clouds and employs a customized point encoder, called Point Embedding Network, to generate 3D geometric embeddings independently. These produced embeddings are then concatenated and processed by our proposed spatial-aware action expert, called 3D-enhanced Action Expert, which combines information from different sensor modalities to produce precise action sequences. Through extensive experiments in both simulation and real-world environments, GeoVLA demonstrates superior performance and robustness. It achieves state-of-the-art results in the LIBERO and ManiSkill2 simulation benchmarks and shows remarkable robustness in real-world tasks requiring height adaptability, scale awareness and viewpoint invariance.

cs.RO↗

arXiv2025

OpenSeg-R: Improving Open-Vocabulary Segmentation via Step-by-Step Visual Reasoning

Open-Vocabulary Segmentation (OVS) has drawn increasing attention for its capacity to generalize segmentation beyond predefined categories. However, existing methods typically predict segmentation masks with simple forward inference, lacking explicit reasoning and interpretability. This makes it challenging for OVS model to distinguish similar categories in open-world settings due to the lack of contextual understanding and discriminative visual cues. To address this limitation, we propose a step-by-step visual reasoning framework for open-vocabulary segmentation, named OpenSeg-R. The proposed OpenSeg-R leverages Large Multimodal Models (LMMs) to perform hierarchical visual reasoning before segmentation. Specifically, we generate both generic and image-specific reasoning for each image, forming structured triplets that explain the visual reason for objects in a coarse-to-fine manner. Based on these reasoning steps, we can compose detailed description prompts, and feed them to the segmentor to produce more accurate segmentation masks. To the best of our knowledge, OpenSeg-R is the first framework to introduce explicit step-by-step visual reasoning into OVS. Experimental results demonstrate that OpenSeg-R significantly outperforms state-of-the-art methods on open-vocabulary semantic segmentation across five benchmark datasets. Moreover, it achieves consistent gains across all metrics on open-vocabulary panoptic segmentation. Qualitative results further highlight the effectiveness of our reasoning-guided framework in improving both segmentation precision and interpretability. Our code is publicly available at https://github.com/Hanzy1996/OpenSeg-R.

cs.CV↗

arXiv2025

SSLFusion: Scale & Space Aligned Latent Fusion Model for Multimodal 3D Object Detection

Multimodal 3D object detection based on deep neural networks has indeed made significant progress. However, it still faces challenges due to the misalignment of scale and spatial information between features extracted from 2D images and those derived from 3D point clouds. Existing methods usually aggregate multimodal features at a single stage. However, leveraging multi-stage cross-modal features is crucial for detecting objects of various scales. Therefore, these methods often struggle to integrate features across different scales and modalities effectively, thereby restricting the accuracy of detection. Additionally, the time-consuming Query-Key-Value-based (QKV-based) cross-attention operations often utilized in existing methods aid in reasoning the location and existence of objects by capturing non-local contexts. However, this approach tends to increase computational complexity. To address these challenges, we present SSLFusion, a novel Scale & Space Aligned Latent Fusion Model, consisting of a scale-aligned fusion strategy (SAF), a 3D-to-2D space alignment module (SAM), and a latent cross-modal fusion module (LFM). SAF mitigates scale misalignment between modalities by aggregating features from both images and point clouds across multiple levels. SAM is designed to reduce the inter-modal gap between features from images and point clouds by incorporating 3D coordinate information into 2D image features. Additionally, LFM captures cross-modal non-local contexts in the latent space without utilizing the QKV-based attention operations, thus mitigating computational complexity. Experiments on the KITTI and DENSE datasets demonstrate that our SSLFusion outperforms state-of-the-art methods. Our approach obtains an absolute gain of 2.15% in 3D AP, compared with the state-of-art method GraphAlign on the moderate level of the KITTI test set.

cs.CV↗

arXiv2025

Glad: A Streaming Scene Generator for Autonomous Driving

The generation and simulation of diverse real-world scenes have significant application value in the field of autonomous driving, especially for the corner cases. Recently, researchers have explored employing neural radiance fields or diffusion models to generate novel views or synthetic data under driving scenes. However, these approaches suffer from unseen scenes or restricted video length, thus lacking sufficient adaptability for data generation and simulation. To address these issues, we propose a simple yet effective framework, named Glad, to generate video data in a frame-by-frame style. To ensure the temporal consistency of synthetic video, we introduce a latent variable propagation module, which views the latent features of previous frame as noise prior and injects it into the latent features of current frame. In addition, we design a streaming data sampler to orderly sample the original image in a video clip at continuous iterations. Given the reference frame, our Glad can be viewed as a streaming simulator by generating the videos for specific scenes. Extensive experiments are performed on the widely-used nuScenes dataset. Experimental results demonstrate that our proposed Glad achieves promising performance, serving as a strong baseline for online video generation. We will release the source code and models publicly.

cs.RO↗

arXiv2024

CLIPer: Hierarchically Improving Spatial Representation of CLIP for Open-Vocabulary Semantic Segmentation

Contrastive Language-Image Pre-training (CLIP) exhibits strong zero-shot classification ability on various image-level tasks, leading to the research to adapt CLIP for pixel-level open-vocabulary semantic segmentation without additional training. The key is to improve spatial representation of image-level CLIP, such as replacing self-attention map at last layer with self-self attention map or vision foundation model based attention map. In this paper, we present a novel hierarchical framework, named CLIPer, that hierarchically improves spatial representation of CLIP. The proposed CLIPer includes an early-layer fusion module and a fine-grained compensation module. We observe that, the embeddings and attention maps at early layers can preserve spatial structural information. Inspired by this, we design the early-layer fusion module to generate segmentation map with better spatial coherence. Afterwards, we employ a fine-grained compensation module to compensate the local details using the self-attention maps of diffusion model. We conduct the experiments on seven segmentation datasets. Our proposed CLIPer achieves the state-of-the-art performance on these datasets. For instance, using ViT-L, CLIPer has the mIoU of 69.8% and 43.3% on VOC and COCO Object, outperforming ProxyCLIP by 9.2% and 4.1% respectively.

cs.CV↗

arXiv2024

VideoGLaMM: A Large Multimodal Model for Pixel-Level Visual Grounding in Videos

Fine-grained alignment between videos and text is challenging due to complex spatial and temporal dynamics in videos. Existing video-based Large Multimodal Models (LMMs) handle basic conversations but struggle with precise pixel-level grounding in videos. To address this, we introduce VideoGLaMM, a LMM designed for fine-grained pixel-level grounding in videos based on user-provided textual inputs. Our design seamlessly connects three key components: a Large Language Model, a dual vision encoder that emphasizes both spatial and temporal details, and a spatio-temporal decoder for accurate mask generation. This connection is facilitated via tunable V-L and L-V adapters that enable close Vision-Language (VL) alignment. The architecture is trained to synchronize both spatial and temporal elements of video content with textual instructions. To enable fine-grained grounding, we curate a multimodal dataset featuring detailed visually-grounded conversations using a semiautomatic annotation pipeline, resulting in a diverse set of 38k video-QA triplets along with 83k objects and 671k masks. We evaluate VideoGLaMM on three challenging tasks: Grounded Conversation Generation, Visual Grounding, and Referring Video Segmentation. Experimental results show that our model consistently outperforms existing approaches across all three tasks.

cs.CV↗

arXiv2024

DB-SAM: Delving into High Quality Universal Medical Image Segmentation

Recently, the Segment Anything Model (SAM) has demonstrated promising segmentation capabilities in a variety of downstream segmentation tasks. However in the context of universal medical image segmentation there exists a notable performance discrepancy when directly applying SAM due to the domain gap between natural and 2D/3D medical data. In this work, we propose a dual-branch adapted SAM framework, named DB-SAM, that strives to effectively bridge this domain gap. Our dual-branch adapted SAM contains two branches in parallel: a ViT branch and a convolution branch. The ViT branch incorporates a learnable channel attention block after each frozen attention block, which captures domain-specific local features. On the other hand, the convolution branch employs a light-weight convolutional block to extract domain-specific shallow features from the input medical image. To perform cross-branch feature fusion, we design a bilateral cross-attention block and a ViT convolution fusion block, which dynamically combine diverse information of two branches for mask decoder. Extensive experiments on large-scale medical image dataset with various 3D and 2D medical segmentation tasks reveal the merits of our proposed contributions. On 21 3D medical image segmentation tasks, our proposed DB-SAM achieves an absolute gain of 8.8%, compared to a recent medical SAM adapter in the literature. The code and model are available at https://github.com/AlfredQin/DB-SAM.

eess.IV↗

arXiv2024

iSeg: An Iterative Refinement-based Framework for Training-free Segmentation

Stable diffusion has demonstrated strong image synthesis ability to given text descriptions, suggesting it to contain strong semantic clue for grouping objects. The researchers have explored employing stable diffusion for training-free segmentation. Most existing approaches refine cross-attention map by self-attention map once, demonstrating that self-attention map contains useful semantic information to improve segmentation. To fully utilize self-attention map, we present a deep experimental analysis on iteratively refining cross-attention map with self-attention map, and propose an effective iterative refinement framework for training-free segmentation, named iSeg. The proposed iSeg introduces an entropy-reduced self-attention module that utilizes a gradient descent scheme to reduce the entropy of self-attention map, thereby suppressing the weak responses corresponding to irrelevant global information. Leveraging the entropy-reduced self-attention module, our iSeg stably improves refined cross-attention map with iterative refinement. Further, we design a category-enhanced cross-attention module to generate accurate cross-attention map, providing a better initial input for iterative refinement. Extensive experiments across different datasets and diverse segmentation tasks reveal the merits of proposed contributions, leading to promising performance on diverse segmentation tasks. For unsupervised semantic segmentation on Cityscapes, our iSeg achieves an absolute gain of 3.8% in terms of mIoU compared to the best existing training-free approach in literature. Moreover, our proposed iSeg can support segmentation with different kinds of images and interactions. The project is available at https://linsun449.github.io/iSeg.

cs.CV↗

arXiv2024

Implicit and Explicit Language Guidance for Diffusion-based Visual Perception

Text-to-image diffusion models have shown powerful ability on conditional image synthesis. With large-scale vision-language pre-training, diffusion models are able to generate high-quality images with rich texture and reasonable structure under different text prompts. However, it is an open problem to adapt the pre-trained diffusion model for visual perception. In this paper, we propose an implicit and explicit language guidance framework for diffusion-based perception, named IEDP. Our IEDP comprises an implicit language guidance branch and an explicit language guidance branch. The implicit branch employs frozen CLIP image encoder to directly generate implicit text embeddings that are fed to diffusion model, without using explicit text prompts. The explicit branch utilizes the ground-truth labels of corresponding images as text prompts to condition feature extraction of diffusion model. During training, we jointly train diffusion model by sharing the model weights of these two branches. As a result, implicit and explicit branches can jointly guide feature learning. During inference, we only employ implicit branch for final prediction, which does not require any ground-truth labels. Experiments are performed on two typical perception tasks, including semantic segmentation and depth estimation. Our IEDP achieves promising performance on both tasks. For semantic segmentation, our IEDP has the mIoU$^\text{ss}$ score of 55.9% on AD20K validation set, which outperforms the baseline method VPD by 2.2%. For depth estimation, our IEDP outperforms the baseline method VPD with a relative gain of 11.0%.

cs.CV↗

arXiv2024

Parameter-Efficient Fine-Tuning for Continual Learning: A Neural Tangent Kernel Perspective

Parameter-efficient fine-tuning for continual learning (PEFT-CL) has shown promise in adapting pre-trained models to sequential tasks while mitigating catastrophic forgetting problem. However, understanding the mechanisms that dictate continual performance in this paradigm remains elusive. To unravel this mystery, we undertake a rigorous analysis of PEFT-CL dynamics to derive relevant metrics for continual scenarios using Neural Tangent Kernel (NTK) theory. With the aid of NTK as a mathematical analysis tool, we recast the challenge of test-time forgetting into the quantifiable generalization gaps during training, identifying three key factors that influence these gaps and the performance of PEFT-CL: training sample size, task-level feature orthogonality, and regularization. To address these challenges, we introduce NTK-CL, a novel framework that eliminates task-specific parameter storage while adaptively generating task-relevant features. Aligning with theoretical guidance, NTK-CL triples the feature representation of each sample, theoretically and empirically reducing the magnitude of both task-interplay and task-specific generalization gaps. Grounded in NTK analysis, our framework imposes an adaptive exponential moving average mechanism and constraints on task-level feature orthogonality, maintaining intra-task NTK forms while attenuating inter-task NTK forms. Ultimately, by fine-tuning optimizable parameters with appropriate regularization, NTK-CL achieves state-of-the-art performance on established PEFT-CL benchmarks. This work provides a theoretical foundation for understanding and improving PEFT-CL models, offering insights into the interplay between feature representation, task orthogonality, and generalization, contributing to the development of more efficient continual learning systems.

cs.LG↗

arXiv2024

Multi-Granularity Language-Guided Training for Multi-Object Tracking

Most existing multi-object tracking methods typically learn visual tracking features via maximizing dis-similarities of different instances and minimizing similarities of the same instance. While such a feature learning scheme achieves promising performance, learning discriminative features solely based on visual information is challenging especially in case of environmental interference such as occlusion, blur and domain variance. In this work, we argue that multi-modal language-driven features provide complementary information to classical visual features, thereby aiding in improving the robustness to such environmental interference. To this end, we propose a new multi-object tracking framework, named LG-MOT, that explicitly leverages language information at different levels of granularity (scene-and instance-level) and combines it with standard visual features to obtain discriminative representations. To develop LG-MOT, we annotate existing MOT datasets with scene-and instance-level language descriptions. We then encode both instance-and scene-level language information into high-dimensional embeddings, which are utilized to guide the visual features during training. At inference, our LG-MOT uses the standard visual features without relying on annotated language descriptions. Extensive experiments on three benchmarks, MOT17, DanceTrack and SportsMOT, reveal the merits of the proposed contributions leading to state-of-the-art performance. On the DanceTrack test set, our LG-MOT achieves an absolute gain of 2.2\% in terms of target object association (IDF1 score), compared to the baseline using only visual features. Further, our LG-MOT exhibits strong cross-domain generalizability. The dataset and code will be available at https://github.com/WesLee88524/LG-MOT.

cs.CV↗

arXiv2024

Vision Foundation Model Driven Foreground-Aware Pseudo-LiDAR Generation for Monocular 3D Object Detection

Pseudo-LiDAR has become a promising paradigm for monocular 3D object detection by transforming monocular images into point cloud representations that can be processed by LiDAR-based 3D object detectors. Recent vision foundation models provide powerful geometric and semantic priors, creating new opportunities for improving the quality of pseudo-LiDAR generation. However, effectively exploiting these priors to produce reliable pseudo-LiDAR remains challenging due to inaccurate depth estimation, insufficient foreground awareness, and redundant background points.In this paper, we propose VFMM3D, a vision foundation model-driven framework for foreground-aware pseudo-LiDAR generation. VFMM3D leverages the depth prior provided by the Depth Anything Model (DAM) and the foreground prior provided by the Segment Anything Model (SAM). A foreground-aware pseudo-LiDAR painting operation is introduced to associate projected 3D points with object-level foreground information, thereby strengthening object structures and suppressing irrelevant background regions. Furthermore, a sparsification strategy is developed to remove redundant points, reduce computational overhead, and improve the compatibility of the generated point clouds with LiDAR-based 3D object detectors. Comprehensive experiments are conducted on two challenging 3D object detection datasets, KITTI and Waymo. Our VFMM3D establishes a new state-of-the-art performance on both datasets. Additionally, experimental results demonstrate the generality of VFMM3D, showcasing its seamless integration into various LiDAR-based 3D object detectors.

cs.CV↗

arXiv2024

SGD: Street View Synthesis with Gaussian Splatting and Diffusion Prior

Novel View Synthesis (NVS) for street scenes play a critical role in the autonomous driving simulation. The current mainstream technique to achieve it is neural rendering, such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS). Although thrilling progress has been made, when handling street scenes, current methods struggle to maintain rendering quality at the viewpoint that deviates significantly from the training viewpoints. This issue stems from the sparse training views captured by a fixed camera on a moving vehicle. To tackle this problem, we propose a novel approach that enhances the capacity of 3DGS by leveraging prior from a Diffusion Model along with complementary multi-modal data. Specifically, we first fine-tune a Diffusion Model by adding images from adjacent frames as condition, meanwhile exploiting depth data from LiDAR point clouds to supply additional spatial information. Then we apply the Diffusion Model to regularize the 3DGS at unseen views during training. Experimental results validate the effectiveness of our method compared with current state-of-the-art models, and demonstrate its advance in rendering images from broader views.

cs.CV↗

arXiv2024

CLIP-VIS: Adapting CLIP for Open-Vocabulary Video Instance Segmentation

Open-vocabulary video instance segmentation strives to segment and track instances belonging to an open set of categories in a videos. The vision-language model Contrastive Language-Image Pre-training (CLIP) has shown robust zero-shot classification ability in image-level open-vocabulary tasks. In this paper, we propose a simple encoder-decoder network, called CLIP-VIS, to adapt CLIP for open-vocabulary video instance segmentation. Our CLIP-VIS adopts frozen CLIP and introduces three modules, including class-agnostic mask generation, temporal topK-enhanced matching, and weighted open-vocabulary classification. Given a set of initial queries, class-agnostic mask generation introduces a pixel decoder and a transformer decoder on CLIP pre-trained image encoder to predict query masks and corresponding object scores and mask IoU scores. Then, temporal topK-enhanced matching performs query matching across frames using the K mostly matched frames. Finally, weighted open-vocabulary classification first employs mask pooling to generate query visual features from CLIP pre-trained image encoder, and second performs weighted classification using object scores and mask IoU scores. Our CLIP-VIS does not require the annotations of instance categories and identities. The experiments are performed on various video instance segmentation datasets, which demonstrate the effectiveness of our proposed method, especially for novel categories. When using ConvNeXt-B as backbone, our CLIP-VIS achieves the AP and APn scores of 32.2% and 40.2% on the validation set of LV-VIS dataset, which outperforms OV2Seg by 11.1% and 23.9% respectively. We will release the source code and models at https://github.com/zwq456/CLIP-VIS.git.

cs.CV↗