arXiv:2603.13099v1 Announce Type: new Abstract: We introduce **CRYSTAL** (*__C__lear __R__easoning via __Y__ielded __S__teps, __T__raceability and __L__ogic*), a diagnostic benchmark with 6,372 instances that evaluates multimodal reasoning through verifiable intermediate steps. We propose two complementary metrics: *Match F1*, which scores step-level precision and recall via semantic similarity matching, and *Ordered Match F1*, which further penalizes disordered reasoning chains. References are constructed through a Delphi-inspired pipeline where four independent MLLMs generate trajectories, aggregated via semantic clustering and validated through human quality gates. Evaluation of 20 MLLMs, including commercial frontier systems not used during benchmark construction, reveals systematic failures invisible to accuracy: universal cherry-picking (precision far exceeds recall), non-monotonic scaling trade-offs, and disordered reasoning where no competitive model preserves more than 60% of matched steps in correct order. Beyond evaluation, we propose the **Causal Process Reward (CPR)**, a multiplicative reward that couples answer correctness with step-level alignment, and **CPR-Curriculum**, which progressively increases reasoning difficulty during training. CPR-Curriculum achieves +32% Match F1 via GRPO where additive reward strategies fail, improving reasoning without manual step annotation.

arXiv:2603.12310v1 Announce Type: cross Abstract: Despite rapid advancements in video generation models, aligning their outputs with complex user intent remains challenging. Existing test-time optimization methods are typically either computationally expensive or require white-box access to model internals. To address this, we present VQQA (Video Quality Question Answering), a unified, multi-agent framework generalizable across diverse input modalities and video generation tasks. By dynamically generating visual questions and using the resulting Vision-Language Model (VLM) critiques as semantic gradients, VQQA replaces traditional, passive evaluation metrics with human-interpretable, actionable feedback. This enables a highly efficient, closed-loop prompt optimization process via a black-box natural language interface. Extensive experiments demonstrate that VQQA effectively isolates and resolves visual artifacts, substantially improving generation quality in just a few refinement steps. Applicable to both text-to-video (T2V) and image-to-video (I2V) tasks, our method achieves absolute improvements of +11.57% on T2V-CompBench and +8.43% on VBench2 over vanilla generation, significantly outperforming state-of-the-art stochastic search and prompt optimization techniques.

arXiv:2603.12382v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have advanced from image-level reasoning to pixel-level grounding, but extending these capabilities to videos remains challenging as models must achieve spatial precision and temporally consistent reference tracking. Existing video MLLMs often rely on a static segmentation token ([SEG]) for frame-wise grounding, which provides semantics but lacks temporal context, causing spatial drift, identity switches, and unstable initialization when objects move or reappear. We introduce SPARROW, a pixel-grounded video MLLM that unifies spatial accuracy and temporal stability through two key components: (i) Target-Specific Tracked Features (TSF), which inject temporally aligned referent cues during training, and (ii) a dual-prompt design that decodes box ([BOX]) and segmentation ([SEG]) tokens to fuse geometric priors with semantic grounding. SPARROW is supported by a curated referential video dataset of 30,646 videos and 45,231 Q&A pairs and operates end-to-end without external detectors via a class-agnostic SAM2-based proposer. Integrated into three recent open-source video MLLMs (UniPixel, GLUS, and VideoGLaMM), SPARROW delivers consistent gains across six benchmarks, improving up to +8.9 J&F on RVOS, +5 mIoU on visual grounding, and +5.4 CLAIR on GCG. These results demonstrate that SPARROW substantially improves referential stability, spatial precision, and temporal coherence in pixel-grounded video understanding. Project page: https://risys-lab.github.io/SPARROW

arXiv:2603.12433v1 Announce Type: cross Abstract: Model stitching, connecting early layers of one model (source) to later layers of another (target) via a light stitch layer, has served as a probe of representational compatibility. Prior work finds that models trained on the same dataset remain stitchable (negligible accuracy drop) despite different initializations or objectives. We revisit stitching for Vision Foundation Models (VFMs) that vary in objectives, data, and modality mix (e.g., CLIP, DINOv2, SigLIP 2) and ask: Are heterogeneous VFMs stitchable? We introduce a systematic protocol spanning the stitch points, stitch layer families, training losses, and downstream tasks. Three findings emerge. (1) Stitch layer training matters: conventional approaches that match the intermediate features at the stitch point or optimize the task loss end-to-end struggle to retain accuracy, especially at shallow stitch points. (2) With a simple feature-matching loss at the target model's penultimate layer, heterogeneous VFMs become reliably stitchable across vision tasks. (3) For deep stitch points, the stitched model can surpass either constituent model at only a small inference overhead (for the stitch layer). Building on these findings, we further propose the VFM Stitch Tree (VST), which shares early layers across VFMs while retaining their later layers, yielding a controllable accuracy-latency trade-off for multimodal LLMs that often leverage multiple VFMs. Taken together, our study elevates stitching from a diagnostic probe to a practical recipe for integrating complementary VFM strengths and pinpointing where their representations align or diverge.

arXiv:2603.12445v1 Announce Type: cross Abstract: Convolutional Neural Networks have shown promising effectiveness in identifying different types of cancer from radiographs. However, the opaque nature of CNNs makes it difficult to fully understand the way they operate, limiting their assessment to empirical evaluation. Here we study the soundness of the standard practices by which CNNs are evaluated for the purpose of cancer pathology. Thirteen highly used cancer benchmark datasets were analyzed, using four common CNN architectures and different types of cancer, such as melanoma, carcinoma, colorectal cancer, and lung cancer. We compared the accuracy of each model with that of datasets made of cropped segments from the background of the original images that do not contain clinically relevant content. Because the rendered datasets contain no clinical information, the null hypothesis is that the CNNs should provide mere chance-based accuracy when classifying these datasets. The results show that the CNN models provided high accuracy when using the cropped segments, sometimes as high as 93\%, even though they lacked biomedical information. These results show that some CNN architectures are more sensitive to bias than others. The analysis shows that the common practices of machine learning evaluation might lead to unreliable results when applied to cancer pathology. These biases are very difficult to identify, and might mislead researchers as they use available benchmark datasets to test the efficacy of CNN methods.

arXiv:2603.12506v1 Announce Type: cross Abstract: Text-to-Image (T2I) generation is primarily driven by Diffusion Models (DM) which rely on random Gaussian noise. Thus, like playing the slots at a casino, a DM will produce different results given the same user-defined inputs. This imposes a gambler's burden: To perform multiple generation cycles to obtain a satisfactory result. However, even though DMs use stochastic sampling to seed generation, the distribution of generated content quality highly depends on the prompt and the generative ability of a DM with respect to it. To account for this, we propose Na\"ive PAINE for improving the generative quality of Diffusion Models by leveraging T2I preference benchmarks. We directly predict the numerical quality of an image from the initial noise and given prompt. Na\"ive PAINE then selects a handful of quality noises and forwards them to the DM for generation. Further, Na\"ive PAINE provides feedback on the DM generative quality given the prompt and is lightweight enough to seamlessly fit into existing DM pipelines. Experimental results demonstrate that Na\"ive PAINE outperforms existing approaches on several prompt corpus benchmarks.

arXiv:2603.12538v1 Announce Type: cross Abstract: Referring image segmentation aims to produce a pixel-level mask for the image region described by a natural-language expression. Although pretrained vision-language models have improved semantic grounding, many existing methods still rely on uniform refinement strategies that do not fully match the diverse reasoning requirements of referring expressions. Because of this mismatch, predictions often contain fragmented regions, inaccurate boundaries, or even the wrong object, especially when pretrained backbones are frozen for computational efficiency. To address these limitations, we propose SERA, a Spatio-Semantic Expert Routing Architecture for referring image segmentation. SERA introduces lightweight, expression-aware expert refinement at two complementary stages within a vision-language framework. First, we design SERA-Adapter, which inserts an expression-conditioned adapter into selected backbone blocks to improve spatial coherence and boundary precision through expert-guided refinement and cross-modal attention. We then introduce SERA-Fusion, which strengthens intermediate visual representations by reshaping token features into spatial grids and applying geometry-preserving expert transformations before multimodal interaction. In addition, a lightweight routing mechanism adaptively weights expert contributions while remaining compatible with pretrained representations. To make this routing stable under frozen encoders, SERA uses a parameter-efficient tuning strategy that updates only normalization and bias terms, affecting less than 1% of the backbone parameters. Experiments on standard referring image segmentation benchmarks show that SERA consistently outperforms strong baselines, with especially clear gains on expressions that require accurate spatial localization and precise boundary delineation.

arXiv:2603.12581v1 Announce Type: cross Abstract: Although diffusion models have achieved remarkable progress in multi-modal magnetic resonance imaging (MRI) translation tasks, existing methods still tend to suffer from anatomical inconsistencies or degraded texture details when handling arbitrary missing-modality scenarios. To address these issues, we propose a latent diffusion-based multi-modal MRI translation framework, termed MSG-LDM. By leveraging the available modalities, the proposed method infers complete structural information, which preserves reliable boundary details. Specifically, we introduce a style--structure disentanglement mechanism in the latent space, which explicitly separates modality-specific style features from shared structural representations, and jointly models low-frequency anatomical layouts and high-frequency boundary details in a multi-scale feature space. During the structure disentanglement stage, high-frequency structural information is explicitly incorporated to enhance feature representations, guiding the model to focus on fine-grained structural cues while learning modality-invariant low-frequency anatomical representations. Furthermore, to reduce interference from modality-specific styles and improve the stability of structure representations, we design a style consistency loss and a structure-aware loss. Extensive experiments on the BraTS2020 and WMH datasets demonstrate that the proposed method outperforms existing MRI synthesis approaches, particularly in reconstructing complete structures. The source code is publicly available at https://github.com/ziyi-start/MSG-LDM.

arXiv:2603.12606v1 Announce Type: cross Abstract: Current vision-language detection and grounding models predominantly focus on prompts with positive semantics and often struggle to accurately interpret and ground complex expressions containing negative semantics. A key reason for this limitation is the lack of high-quality training data that explicitly captures discriminative negative samples and negation-aware language descriptions. To address this challenge, we introduce D-Negation, a new dataset that provides objects annotated with both positive and negative semantic descriptions. Building upon the observation that negation reasoning frequently appears in natural language, we further propose a grouped opposition-based learning framework that learns negation-aware representations from limited samples. Specifically, our method organizes opposing semantic descriptions from D-Negation into structured groups and formulates two complementary loss functions that encourage the model to reason about negation and semantic qualifiers. We integrate the proposed dataset and learning strategy into a state-of-the-art language-based grounding model. By fine-tuning fewer than 10 percent of the model parameters, our approach achieves improvements of up to 4.4 mAP and 5.7 mAP on positive and negative semantic evaluations, respectively. These results demonstrate that explicitly modeling negation semantics can substantially enhance the robustness and localization accuracy of vision-language grounding models.

arXiv:2603.12625v1 Announce Type: cross Abstract: Multimodal recommendation is commonly framed as a feature fusion problem, where textual and visual signals are combined to better model user preference. However, the effectiveness of multimodal recommendation may depend not only on how modalities are fused, but also on whether item content is represented in a semantic space aligned with preference matching. This issue is particularly important because raw visual features often preserve appearance similarity, while user decisions are typically driven by higher-level semantic factors such as style, material, and usage context. Motivated by this observation, we propose LVLM-grounded Multimodal Semantic Representation for Recommendation (VLM4Rec), a lightweight framework that organizes multimodal item content through semantic alignment rather than direct feature fusion. VLM4Rec first uses a large vision-language model to ground each item image into an explicit natural-language description, and then encodes the grounded semantics into dense item representations for preference-oriented retrieval. Recommendation is subsequently performed through a simple profile-based semantic matching mechanism over historical item embeddings, yielding a practical offline-online decomposition. Extensive experiments on multiple multimodal recommendation datasets show that VLM4Rec consistently improves performance over raw visual features and several fusion-based alternatives, suggesting that representation quality may matter more than fusion complexity in this setting. The code is released at https://github.com/tyvalencia/enhancing-mm-rec-sys.

arXiv:2603.12647v1 Announce Type: cross Abstract: Recent 3D Gaussian Splatting (3DGS) methods have demonstrated the feasibility of self-driving scene reconstruction and novel view synthesis. However, most existing methods either rely solely on cameras or use LiDAR only for Gaussian initialization or depth supervision, while the rich scene information contained in point clouds, such as reflectance, and the complementarity between LiDAR and RGB have not been fully exploited, leading to degradation in challenging self-driving scenes, such as those with high ego-motion and complex lighting. To address these issues, we propose a robust and efficient LiDAR-reflectance-guided Salient Gaussian Splatting method (LR-SGS) for self-driving scenes, which introduces a structure-aware Salient Gaussian representation, initialized from geometric and reflectance feature points extracted from LiDAR and refined through a salient transform and improved density control to capture edge and planar structures. Furthermore, we calibrate LiDAR intensity into reflectance and attach it to each Gaussian as a lighting-invariant material channel, jointly aligned with RGB to enforce boundary consistency. Extensive experiments on the Waymo Open Dataset demonstrate that LR-SGS achieves superior reconstruction performance with fewer Gaussians and shorter training time. In particular, on Complex Lighting scenes, our method surpasses OmniRe by 1.18 dB PSNR.

arXiv:2603.12667v1 Announce Type: cross Abstract: Aggregates, serving as the main skeleton in assemblies of construction materials, are important functional components in various building and transportation infrastructures. They can be used in unbound layer applications, e.g. pavement base and railroad ballast, bound applications of cement concrete and asphalt concrete, and as riprap and large-sized primary crushed rocks. Information on the size and shape or morphology of aggregates can greatly facilitate the Quality Assurance/Quality Control (QA/QC) process by providing insights of aggregate behavior during composition and packing. A full 3D characterization of aggregate particle morphology is difficult both during production in a quarry and at a construction site. Many aggregate imaging approaches have been developed to quantify the particle morphology by computer vision, including 2D image-based approaches that analyze particle silhouettes and 3D scanning-based methods that require expensive devices such as 3D laser scanners or X-Ray Computed Tomography (CT) equipment. This paper presents a flexible and cost-effective photogrammetry-based approach for the 3D reconstruction of aggregate particles. The proposed approach follows a marker-based design that enables background suppression, point cloud stitching, and scale referencing to obtain high-quality aggregate models. The accuracy of the reconstruction results was validated against ground-truth for selected aggregate samples. Comparative analyses were conducted on 2D and 3D morphological properties of the selected samples. Significant differences were found between the 2D and 3D statistics. Based on the presented approach, 3D shape information of aggregates can be obtained easily and at a low cost, thus allowing convenient aggregate inspection, data collection, and 3D morphological analysis.

arXiv:2603.12693v1 Announce Type: cross Abstract: This article presents our results for the 10th Affective Behavior Analysis in-the-Wild (ABAW) competition. For frame-wise facial emotion understanding tasks (frame-wise facial expression recognition, valence-arousal estimation, action unit detection), we propose a fast approach based on facial embedding extraction with pre-trained EfficientNet-based emotion recognition models. If the latter model's confidence exceeds a threshold, its prediction is used. Otherwise, we feed embeddings into a simple multi-layered perceptron trained on the AffWild2 dataset. Estimated class-level scores are smoothed in a sliding window of fixed size to mitigate noise in frame-wise predictions. For the fine-grained violence detection task, we examine several pre-trained architectures for frame embeddings and their aggregation for video classification. Experimental results on four tasks from the ABAW challenge demonstrate that our approach significantly improves validation metrics over existing baselines.

arXiv:2603.12719v1 Announce Type: cross Abstract: Point cloud registration (PCR) is a fundamental task in 3D vision and provides essential support for applications such as autonomous driving, robotics, and environmental modeling. Despite its widespread use, existing methods often fail when facing real-world challenges like heavy noise, significant occlusions, and large-scale transformations. These limitations frequently result in compromised registration accuracy and insufficient robustness in complex environments. In this paper, we propose IGASA as a novel registration framework constructed upon a Hierarchical Pyramid Architecture (HPA) designed for robust multi-scale feature extraction and fusion. The framework integrates two pivotal components consisting of the Hierarchical Cross-Layer Attention (HCLA) module and the Iterative Geometry-Aware Refinement (IGAR) module. The HCLA module utilizes skip attention mechanisms to align multi-resolution features and enhance local geometric consistency. Simultaneously, the IGAR module is designed for the fine matching phase by leveraging reliable correspondences established during coarse matching. This synergistic integration within the architecture allows IGASA to adapt effectively to diverse point cloud structures and intricate transformations. We evaluate the performance of IGASA on four widely recognized benchmark datasets including 3D(Lo)Match, KITTI, and nuScenes. Our extensive experiments consistently demonstrate that IGASA significantly surpasses state-of-the-art methods and achieves notable improvements in registration accuracy. This work provides a robust foundation for advancing point cloud registration techniques while offering valuable insights for practical 3D vision applications. The code for IGASA is available in \href{https://github.com/DongXu-Zhang/IGASA}{https://github.com/DongXu-Zhang/IGASA}.

arXiv:2603.12721v1 Announce Type: cross Abstract: Robust point cloud registration is a fundamental task in 3D computer vision and geometric deep learning, essential for applications such as large-scale 3D reconstruction, augmented reality, and scene understanding. However, the performance of established learning-based methods often degrades in complex, real world scenarios characterized by incomplete data, sensor noise, and low overlap regions. To address these limitations, we propose CMHANet, a novel Cross-Modal Hybrid Attention Network. Our method integrates the fusion of rich contextual information from 2D images with the geometric detail of 3D point clouds, yielding a comprehensive and resilient feature representation. Furthermore, we introduce an innovative optimization function based on contrastive learning, which enforces geometric consistency and significantly improves the model's robustness to noise and partial observations. We evaluated CMHANet on the 3DMatch and the challenging 3DLoMatch datasets. \rev{Additionally, zero-shot evaluations on the TUM RGB-D SLAM dataset verify the model's generalization capability to unseen domains.} The experimental results demonstrate that our method achieves substantial improvements in both registration accuracy and overall robustness, outperforming current techniques. We also release our code in \href{https://github.com/DongXu-Zhang/CMHANet}{https://github.com/DongXu-Zhang/CMHANet}.

arXiv:2603.12722v1 Announce Type: cross Abstract: Visual stimuli reconstruction from EEG remains challenging due to fidelity loss and representation shift. We propose CognitionCapturerPro, an enhanced framework that integrates EEG with multi-modal priors (images, text, depth, and edges) via collaborative training. Our core contributions include an uncertainty-weighted similarity scoring mechanism to quantify modality-specific fidelity and a fusion encoder for integrating shared representations. By employing a simplified alignment module and a pre-trained diffusion model, our method significantly outperforms the original CognitionCapturer on the THINGS-EEG dataset, improving Top-1 and Top-5 retrieval accuracy by 25.9% and 10.6%, respectively. Code is available at: https://github.com/XiaoZhangYES/CognitionCapturerPro.

arXiv:2603.12743v1 Announce Type: cross Abstract: Concept customization typically binds rare tokens to a target concept. Unfortunately, these approaches often suffer from unstable performance as the pretraining data seldom contains these rare tokens. Meanwhile, these rare tokens fail to convey the inherent knowledge of the target concept. Consequently, we introduce Knowledge-aware Concept Customization, a novel task aiming at binding diverse textual knowledge to target visual concepts. This task requires the model to identify the knowledge within the text prompt to perform high-fidelity customized generation. Meanwhile, the model should efficiently bind all the textual knowledge to the target concept. Therefore, we propose MoKus, a novel framework for knowledge-aware concept customization. Our framework relies on a key observation: cross-modal knowledge transfer, where modifying knowledge within the text modality naturally transfers to the visual modality during generation. Inspired by this observation, MoKus contains two stages: (1) In visual concept learning, we first learn the anchor representation to store the visual information of the target concept. (2) In textual knowledge updating, we update the answer for the knowledge queries to the anchor representation, enabling high-fidelity customized generation. To further comprehensively evaluate our proposed MoKus on the new task, we introduce the first benchmark for knowledge-aware concept customization: KnowCusBench. Extensive evaluations have demonstrated that MoKus outperforms state-of-the-art methods. Moreover, the cross-model knowledge transfer allows MoKus to be easily extended to other knowledge-aware applications like virtual concept creation and concept erasure. We also demonstrate the capability of our method to achieve improvements on world knowledge benchmarks.

arXiv:2603.12758v1 Announce Type: cross Abstract: Reliable multi-object tracking (MOT) is essential for robotic systems operating in complex and dynamic environments. Despite recent advances in detection and association, online MOT methods remain vulnerable to identity switches caused by frequent occlusions and object overlap, where incorrect associations can propagate over time and degrade tracking reliability. We present a lightweight post-association correction framework (FC-Track) for online MOT that explicitly targets overlap-induced mismatches during inference. The proposed method suppresses unreliable appearance updates under high-overlap conditions using an Intersection over Area (IoA)-based filtering strategy, and locally corrects detection-to-tracklet mismatches through appearance similarity comparison within overlapped tracklet pairs. By preventing short-term mismatches from propagating, our framework effectively mitigates long-term identity switches without resorting to global optimization or re-identification. The framework operates online without global optimization or re-identification, making it suitable for real-time robotic applications. We achieve 81.73 MOTA, 82.81 IDF1, and 66.95 HOTA on the MOT17 test set with a running speed of 5.7 FPS, and 77.52 MOTA, 80.90 IDF1, and 65.67 HOTA on the MOT20 test set with a running speed of 0.6 FPS. Specifically, our framework FC-Track produces only 29.55% long-term identity switches, which is substantially lower than existing online trackers. Meanwhile, our framework maintains state-of-the-art performance on the MOT20 benchmark.

arXiv:2603.12773v1 Announce Type: cross Abstract: In recent years, learning-based underwater image enhancement (UIE) techniques have rapidly evolved. However, distribution shifts between high-quality enhanced outputs and natural images can hinder semantic cue extraction for downstream vision tasks, thereby limiting the adaptability of existing enhancement models. To address this challenge, this work proposes a new learning mechanism that leverages Vision-Language Models (VLMs) to empower UIE models with semantic-sensitive capabilities. To be concrete, our strategy first generates textual descriptions of key objects from a degraded image via VLMs. Subsequently, a text-image alignment model remaps these relevant descriptions back onto the image to produce a spatial semantic guidance map. This map then steers the UIE network through a dual-guidance mechanism, which combines cross-attention and an explicit alignment loss. This forces the network to focus its restorative power on semantic-sensitive regions during image reconstruction, rather than pursuing a globally uniform improvement, thereby ensuring the faithful restoration of key object features. Experiments confirm that when our strategy is applied to different UIE baselines, significantly boosts their performance on perceptual quality metrics as well as enhances their performance on detection and segmentation tasks, validating its effectiveness and adaptability.

arXiv:2603.12793v1 Announce Type: cross Abstract: A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes and visual representations, making it non-trivial to jointly optimize within a shared feature space. In this work, we present Cheers, a unified multimodal model that decouples patch-level details from semantic representations, thereby stabilizing semantics for multimodal understanding and improving fidelity for image generation via gated detail residuals. Cheers includes three key components: (i) a unified vision tokenizer that encodes and compresses image latent states into semantic tokens for efficient LLM conditioning, (ii) an LLM-based Transformer that unifies autoregressive decoding for text generation and diffusion decoding for image generation, and (iii) a cascaded flow matching head that decodes visual semantics first and then injects semantically gated detail residuals from the vision tokenizer to refine high-frequency content. Experiments on popular benchmarks demonstrate that Cheers matches or surpasses advanced UMMs in both visual understanding and generation. Cheers also achieves 4x token compression, enabling more efficient high-resolution image encoding and generation. Notably, Cheers outperforms the Tar-1.5B on the popular benchmarks GenEval and MMBench, while requiring only 20% of the training cost, indicating effective and efficient (i.e., 4x token compression) unified multimodal modeling. We will release all code and data for future research.