paperlist.ai

AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

MobiCom 2026

Detecting Liquid Food Adulteration with Mobile Hyperspectral Analysis based on RGB Camera

Zimo Liao, Qianyi Huang, Lei Yang

Liquid adulteration poses critical societal challenges by endangering public health and causing economic losses. Existing methods on mobile devices rely solely on vibration-based solutions, which face practical limitations due to requirements of specific container materials/structures and vibrational contact. In this paper, we present LiquidLens, a practical, non-contact mobile system for detecting liquid adulteration based on RGB images of liquid surfaces. Inspired by hyperspectral reconstruction, we aim to analyze liquid food's hyperspectral properties from RGB data for effective adulteration detection. To achieve this, we address three major challenges: First, the specular reflections of liquid, which act as a mirror image of surrounding obstacles, distort spectral measurements. Second, the camera's spectral range is restricted to the visible domain, which lacks access to vital information in the invisible bands. Third, the target liquid spectrum is entangled with uncertain ambient illumination, rendering it uninterpretable. To overcome these challenges, we first exploit the camera flash as visible light source and validate its spectral consistency across mobile devices. Next, we design a specular noise suppression technique to cancel the specular noise. Then, we design an RGB-based VIS-NIR spectral recovery (RVN-SR) framework to overcome the spectral range limitation. Besides, a cross-spectral training scheme is designed to train the RVN-SR model in the flash domain, compensating for the lack of NIR radiation. We deploy and validate LiquidLens across multiple smartphones. Experimental results show robust detection accuracy of 96.2% for 8 honey adulteration types and 94.0% for 5 wine adulteration types.

MobiCom 2026

Spatially Selective Radar Backscatter Communication

Rong Ding, Haiming Jin, Ruoyan Wang, Zijie Chen, Guiyun Fan, Feng Shen, Lu Su, Xiaohua Tian, Linghe Kong, Guihai Chen

This paper presents Serena, the first spatially selective radar backscatter communication system that is capable of delivering messages to radars at specific locations. Serena offers a low-power solution for properly delivering location-specific messages (e.g., no-fly-zone alert, speed-reduction warning, turning notification) to radar-equipped vehicles (e.g., automobile, UAV, UGV). Rather than manipulating RF signals with power-intensive components, Serena leverages the limited modulation capability of multiple low-power backscatter tags to generate spatially selective radar signal feature, and further utilizes such signal feature to represent bits. Such design philosophy is primarily realized by our two core designs: (i) a multi-tag distributed and synchronous modulation technique that generates spatially selective radar signal feature (i.e., the combination of the frequencies of tag-generated harmonics in radar IF spectrum); (ii) an intermittent mutual synchronization protocol that achieves 10-μs level inter-tag synchronization under only 10μW-level power consumption. Our extensive experiments demonstrate that Serena achieves desirable spatial selectivity in diverse environments: it achieves less than 1% average BER at target locations, while keeping the average BER around 30% at non-target locations.

MobiCom 2026

MANA: Towards Efficient Mobile Ad Detection via Multimodal Agentic UI Navigation

Yizhe Zhao, Yongjian Fu, Zihao Feng, Hao Pan, Yongheng Deng, Yaoxue Zhang, Ju Ren

Mobile advertising dominates app monetization but introduces risks ranging from intrusive user experience to malware delivery. Existing detection methods rely either on static analysis, which misses runtime behaviors, or on heuristic UI exploration, which struggles with sparse and obfuscated ads. In this paper, we present MANA, the first agentic multimodal reasoning framework for mobile ad detection. MANA integrates static, visual, temporal, and experiential signals into a reasoning-guided navigation strategy that determines not only how to traverse interfaces but also where to focus, enabling efficient and robust exploration. We implement and evaluate MANA on commercial smartphones over 200 apps, achieving state-of-the-art accuracy and efficiency. Compared to baselines, it improves detection accuracy by 30.5%-56.3% and reduces exploration steps by 29.7%-63.3%. Case studies further demonstrate its ability to uncover obfuscated and malicious ads, underscoring its practicality for mobile ad auditing and its potential for broader runtime UI analysis (e.g., permission abuse). Code and dataset are available at https://github.com/MANA-2026/MANA.

MobiCom 2026

Golden Sniffer: Low-Complexity 5G New Radio Signal Interception Exploiting Gold Sequences

Ivan Palamà, Stefano Mangione, Alessandra Dino, Giulia Focarelli, Giuseppe Bianchi

Sniffing 5G New Radio (NR) transmissions presents a significantly greater challenge compared to LTE, primarily due to the concealment of Radio Network Temporary Identifiers and relevant scrambling parameters. Existing techniques rely on exhaustive searches and simplifying assumptions that often do not hold in real deployments. In this paper, we propose a fundamentally different approach that exploits the linearity and non-cryptographic nature of the Gold sequences used in 5G NR for scrambling, drastically reducing the sniffer parameter space to a handful of test alternatives. We also propose and assess a countermeasure that disrupts the phase-based demodulation reference signal reconstruction used by our approach, with only minor impact on legitimate users.

MobiCom 2026

InstMeter: An Instruction-Level Method to Predict Energy and Latency of DL Model Inference on MCUs

Hao Liu, Qing Wang, Marco Zuniga

Deep learning (DL) models can now run on microcontrollers (MCUs). Through neural architecture search (NAS), we can search DL models that meet the constraints of MCUs. Among various constraints, energy and latency costs of the model inference are critical metrics. To predict them, existing research relies on coarse proxies such as multiply-accumulations (MACs) and model's input parameters, often resulting in inaccurate predictions or requiring extensive data collection. In this paper, we propose InstMeter, a predictor leveraging MCUs' clock cycles to accurately estimate the energy and latency of DL models. Clock cycles are fundamental metrics reflecting MCU operations, directly determining energy and latency costs. Furthermore, a unique property of our predictor is its strong linearity, allowing it to be simple and accurate. We thoroughly evaluate InstMeter under different scenarios, MCUs, and software settings. Compared with state-of-the-art studies, InstMeter can reduce the energy and latency prediction errors by 3× and 6.5×, respectively, while requiring 100× and 10× less training data. In the NAS scenario, InstMeter can fully exploit the energy budget, identifying optimal DL models with higher inference accuracy. We also evaluate InstMeter's generalization performance through various experiments on three ARM MCUs (Cortex-M4, M7, M33) and one RISC-V-based MCU (ESP32-C3), different compilation options (-Os, -O2), GCC versions (v7.3, v10.3), application scenarios (keyword spotting, image recognition), dynamic voltage and frequency scaling, temperatures (21°C, 43°C), and software settings (TFLMv2.4, TFLMvCI). Our source codes and the MCU-specific benchmark datasets are available here: https://github.com/haoliu001/InstMeter.

ECCV 2026

Keep Your Friends Close, and the Right Neighbours Closer: Disaster-Conditioned Kernel-Regularized Graph Attention for Building Damage Classification

Fuad Hasan, Chul Min Yeum

Disaster damage is spatial: buildings rarely fail in isolation.Yet using spatial context for damage classification remains surprisinglyunderexplored, and many pipelines still rely primarily on per-buildingappearance cues even when the dominant uncertainty is spatially struc-tured. Complicating matters, the right neighbourhood is not the sameacross events. Floods, hurricanes, and wildfires can exhibit very differ-ent clustering behaviour, making spatial reasoning valuable but easy tomisuse—naive context aggregation can improve visual coherence whileoversmoothing boundaries or propagating structured errors. We studythis tension on xBD (the dataset used in the xView2 challenge) in a con-trolled post-localization, classification-only setup: each building is repre-sented by a pre/post combined (PPC) patch cropped from the providedpolygons, and spatial context is modelled with GPS-derived buildinggraphs. Our approach keeps local evidence “close” by preserving strongspatial relationships in disaster damage patterns, while bringing onlythe right neighbours “closer” through a disaster-type-conditioned graphmodel that injects a learnable multi-scale spatial kernel prior into atten-tion, allowing the effective neighbourhood scale to adapt across disastertypes rather than being learned as a single global smoothing rule. Todiscourage coherence-by-smoothing, we add a residual de-correlation lossthat penalizes positive Moran’s I in prediction residuals. We evaluate themethod with (i) an xView2 holdout external-reference comparison un-der fixed building instances and (ii) zero-shot transfer evaluations thatstress-test generalization under event and dataset shift: leave-one-event-out (LOEO) on xBD and cross-dataset transfer from xBD to Ida-BD.The model improves macro-F1 and substantially reduces residual spa-tial autocorrelation under zero-shot event shift, indicating better use ofspatial context rather than naive smoothing and enabling more reliabletransfer to unseen events within known disaster types.

ECCV 2026

AeroVLA: A Vision-Language-Action Model for UAV Navigation via Minimalist End-to-End Control

Peng Xu, Zhengnan Deng, Jiayan Deng, Zonghua Gu, Shaohua Wan

Vision-Language Navigation (VLN) for Unmanned AerialVehicles (UAVs) demands complex visual interpretation and continuouscontrol in dynamic 3D environments. Existing hierarchical approachesrely on dense oracle guidance or auxiliary object detectors, creating se-mantic gaps and limiting genuine autonomy. We propose AeroVLA, aminimalist end-to-end Vision-Language-Action framework mapping rawvisual observations and fuzzy linguistic instructions directly to contin-uous physical control signals. First, we introduce a streamlined dual-view perception strategy that reduces visual redundancy while preserv-ing essential cues for forward navigation and precise grounding, whichadditionally facilitates future simulation-to-reality transfer. To reclaimgenuine autonomy, we deploy a fuzzy directional prompting mechanismderived solely from onboard sensors, completely eliminating the depen-dency on dense oracle guidance. Ultimately, we formulate a unified con-trol space that integrates continuous 3-Degree-of-Freedom (3-DoF) kine-matic commands with an intrinsic landing signal, freeing the agent fromexternal object detectors for precision landing. Extensive experiments onthe TravelUAV benchmark demonstrate that AeroVLA achieves state-of-the-art performance in seen environments. Furthermore, it exhibitssuperior generalization in unseen scenarios by achieving nearly threetimes the success rate of leading baselines, validating that a minimal-ist, autonomy-centric paradigm captures more robust visual-motor rep-resentations than complex modular systems. Code is available at: https://github.com/XuPeng23/AeroVLA

ECCV 2026

Agent-OBJ: Prompt-Driven 3D Adversaries for Multi-Modal Perception

Bing Li, Sean Du, Xuhong Ren, Luqi Gong, Wee Peng Tay, Qing Guo

Modern autonomous driving systems increasingly rely ontightly coupled Camera–LiDAR fusion pipelines to achieve robust sceneunderstanding. While this multi-modal redundancy is designed to counterindependent single-sensor failures, it implicitly assumes that physicalobjects cannot simultaneously deceive both modalities. In this paper, weexpose the vulnerability of this assumption by proposing Agent-OBJ,a generative framework that synthesizes physically plausible 3D adver-saries via a prompt-driven agent. Unlike prior works that mainly focus onsensor-specific perturbations, our pipeline instantiates a physically plau-sible 3D adversary by generating a canonical pedestrian geometry from abase prompt and then modulating its shape via an additional semanticprompt, while controlling appearance with a pretrained multi-view per-sonalization model conditioned on multi-view images and a style promptto produce view-consistent appearance. To simultaneously compromiseboth 2D monocular and 3D fusion detectors, the agent dynamically op-timizes the generated adversary using confidence-score feedback fromboth detectors. Meanwhile, to ensure stealthiness, we regularize the gen-erated 3D point cloud by enforcing similarity to a pedestrian featurebank, aligning it with the distribution of natural objects and makingit visually and geometrically indistinguishable from benign instances.Furthermore, we incorporate a multi-view consistency constraint duringoptimization to promote cross-view robustness, ensuring the adversaryremains effective under diverse viewpoints in real-world driving scenes.Extensive experiments on nuScenes demonstrate high attack success ratesagainst both monocular and fusion detectors while preserving strongperceptual plausibility.

ECCV 2026

DVG-WM: Disentangled Video Generation Enables Efficient Embodied World Model for Robotic Manipulation

Ziyu Shan, Zhenyu Wu, Xiaofeng Wang, Zheng Zhu, Ziwei Wang

Video-based embodied world models provide an appealingsubstrate for robotic manipulation by predicting future states, yet cur-rent approaches remain limited by a fundamental entanglement: accu-rately modeling dynamics typically requires low-level temporal reason-ing, while producing high-resolution frames demands expansive visualsynthesis according to high-level semantics. This entanglement results inslow inference speed for iterative planning or too coarse predictions toretain contact-rich details. To solve this dilemma, we present Disentan-gled Video Generation World Model (DVG-WM), an efficient frameworkthat explicitly decomposes world modeling into dynamics learn-ing and visual synthesis. Conditioned on an initial observation anda language instruction, our model first generates a plausible sequence ofintermediate visual states to preview the physical interaction and refinesthem to obtain high-fidelity videos. Furthermore, an efficient cascadingmechanism is proposed, where DVG-WM leverages flow matching to di-rectly map the dynamics to video latents, and introduces a latent degra-dation mechanism to enable the capability of regenerating contact-richdetails. Experiments on LIBERO and real-world platforms demonstrateimproved video quality with up to 3.97 × acceleration, validating thatdisentangled video generation can be an efficient embodied world modelfor robotic manipulation.

ECCV 2026

Boxer: Robust Lifting of Open-World 2D Bounding Boxes to 3D

Daniel DeTone, Tianwei Shen, Fan Zhang, Lingni Ma, Julian Straub, Richard Newcombe, Jakob Engel

Detecting and localizing objects in space is a fundamentalcomputer vision problem. While much progress has been made to solve2D object detection, 3D object localization is much less explored and farfrom solved, especially for open-world categories. To address this researchchallenge, we propose Boxer, an algorithm to estimate static 3D boundingboxes (3DBBs) from 2D open-vocabulary object detections, posed images,and optional depth represented either as a sparse point cloud or densedepth. At its core is BoxerNet, a transformer-based network that lifts2D bounding box (2DBB) proposals into 3D, followed by multi-viewfusion and geometric filtering to produce globally consistent de-duplicated3DBBs in metric world space. Boxer leverages the power of existing 2DBBdetection algorithms (e.g., DETIC [52], OWLv2 [29], SAM3 [8]) to localizeobjects in 2D. This allows the main BoxerNet model to focus on liftingto 3D rather than detecting, ultimately reducing the demand for costlyannotated 3DBB training data. Extending the CuTR [20] formulation, weincorporate an aleatoric uncertainty for robust regression, a mean depthpatch encoding to support sparse depth inputs, and large-scale trainingwith over 1.2 million unique 3DBBs. BoxerNet outperforms state-of-the-art baselines in open-world 3DBB lifting, including CuTR in egocentricsettings without dense depth (0.532 vs. 0.010 mAP) and on CA-1M withdense depth available (0.412 vs. 0.250 mAP). Project page and code areavailable at https://facebookresearch.github.io/boxer.

ECCV 2026

Dual-Anchoring: Addressing State Drift in Vision-Language Navigation

Kangyi Wu, Pengna Li, Kailin Lyu, Xi Lin, Lin Zhao, Qingrong He, Jinjun Wang, Jianyi Liu

Vision-Language Navigation (VLN) requires an agent to navigate through 3D environments by following natural language instructions. While recent Video Large Language Models (Video-LLMs) have largely advanced VLN, they remain highly susceptible to State Drift in long scenarios. In these cases, the agent’s internal state drifts away from the true task execution state, leading to aimless wandering and failure to execute essential maneuvers in the instruction. We attribute this failure to two distinct cognitive deficits: Progress Drift, where the agent fails to distinguish completed sub-goals from remaining ones, and Memory Drift, where the agent’s history representations degrade, making it lose track of visited landmarks. In this paper, we propose a Dual-Anchoring Framework that explicitly anchors the instruction progress and history representations. First, to address progress drift, we introduce Instruction Progress Anchoring, which supervises the agent to generate structured text tokens that delineate completed versus remaining sub-goals. Second, to mitigate memory drift, we propose Memory Landmark Anchoring, which utilizes a Landmark-Centric World Model to retrospectively predict object-centric embeddings extracted by the Segment Anything Model, compelling the agent to explicitly verify past observations and preserve distinct representations of visited landmarks. Facilitating this framework, we curate two extensive datasets: 3.6 million samples with explicit progress descriptions, and 937k grounded landmark data for retrospective verification. Extensive experiments in both simulation and real-world environments demonstrate the superiority of our method, achieving a 15.2% improvement in Success Rate and a remarkable 24.7% gain on long-horizon trajectories. To facilitate further research, we will release our code, data generation pipelines, and the collected datasets.

ECCV 2026

Plug-and-Play Traffic Element Awareness for End-to-End Autonomous Driving

Zongzheng Zhang, Jijun Wang, Saining Zhang, Wang Shuo, Yiru Wang, Hai Yang, Yang Chen, Yuwen Heng, Hao Sun, Jiang anqing, Hao Zhao

Traffic elements such as traffic lights and road signs play a fundamental role in human driving decisions and should naturally influence end-to-end driving performance. However, existing end-to-end driving research predominantly focuses on dynamic road participants (e.g., vehicles and pedestrians), while the role of traffic elements remains largely unexplored. To date, the community lacks a systematic study quantifying how traffic elements affect end-to-end driving models. This gap stems from two main challenges: first, existing public datasets rarely provide structured annotations for traffic elements; second, modern end-to-end driving systems vary widely in architectures and training paradigms, making conclusions drawn from a single method difficult to generalize. In this work, we present the first systematic investigation of traffic element awareness for end-to-end autonomous driving. We begin by constructing a unified research infrastructure by augmenting multiple public driving datasets with comprehensive traffic element annotations. To ensure broad applicability across diverse model families, we intentionally adopt a minimal and universal integration design, allowing traffic element signals to be incorporated into existing pipelines in a plug-and-play manner with negligible architectural modification. We evaluate this design across a wide spectrum of modern paradigms, including perception–prediction-planning pipelines, vision-language-action models (VLA), regression-based planners, diffusion-based policies, and trajectory scoring frameworks. Experiments are conducted on multiple widely used benchmarks, including nuScenes, NAVSIM-v1, NAVSIM-v2, and Bench2Drive. Across all paradigms and datasets, our simple integration consistently improves driving performance, demonstrating that traffic element awareness provides a robust and generalizable signal for end-to-end driving systems. Notably, on the challenging NAVSIM-v2 benchmark, our approach significantly boosts the performance of stateof-the-art architectures and data pipelines, establishing a new state of the art.

ECCV 2026

SEERBench: A Spatial Ego-Exo Reasoning Benchmark for MLLMs with a Simple Yet Effective Baseline

Fengyuan Lu, Jiahe Feng, Zhengyang Zhou, Shaofeng Zhang, Wenbin Li, Qi Fan, Yang Gao

Humans seamlessly integrate egocentric perceptions with exocentric perspectives to comprehend dynamic spatial positioning. For Embodied AI, mastering this ego-exo spatial reasoning is a fundamental prerequisite for real-world interaction. Although Multimodal Large Language Models (MLLMs) have demonstrated significant progress in general visual reasoning, their capacity for ego-exo spatial reasoning in real-world environments remains largely unexplored. To investigate whether current MLLMs exhibit such human-like spatial cognition, we introduce SEERBench, a rigorous benchmark comprising 1,198 highquality, human-annotated question-answer pairs across 35 diverse realworld Ego-Exo scenarios. SEERBench evaluates models through 10 distinct tasks structured into three progressive cognitive levels: Spatial Perception, Spatial Imagination, and Ego-Exo Collaboration. Extensive evaluations of 19 representative MLLMs reveal a substantial performance gap: even the best-performing frontier models trail human experts by a margin of 46.8%, highlighting severe limitations in current multimodal foundations. To address this, we propose SEER-Map, a training-free, tool-augmented baseline. By explicitly constructing top-down Bird’s-Eye View (BEV) spatial topologies to assist in ego-exo geometric alignment, SEER-Map functions as an Oracle-informed probing baseline lower the threshold of spatial understanding. Codes and data are available here.

ECCV 2026

Experts-Guided Unbalanced Optimal Transport for ISP Learning from Unpaired and/or Paired Data

Georgy Perevozchikov, Nancy Mehta, Egor Ershov, Radu Timofte

Learned Image Signal Processing (ISP) pipelines offer powerful end-to-end performance but are critically dependent on large-scale paired raw-to-sRGB datasets. This reliance on costly-to-acquire paired data remains a significant bottleneck. To address this challenge, we introduce a novel, unsupervised training framework based on Optimal Transport capable of training arbitrary ISP architectures in both unpaired and paired modes. We are the first to successfully apply Unbalanced Optimal Transport (UOT) for this complex, cross-domain translation task. Our UOT-based framework provides robustness to outliers in the target sRGB data, allowing it to discount atypical samples that would be prohibitively costly to map. A key component of our framework is a novel “committee of expert discriminators,” a hybrid adversarial regularizer. This committee guides the optimal transport mapping by providing specialized, targeted gradients to correct specific ISP failure modes, including color fidelity, structural artifacts, and frequency-domain realism. To demonstrate the effectiveness of our approach, we retrained existing state-of-the-art ISP architectures using our paired and unpaired setups. Our experiments show that while our framework, when trained in paired mode, improves upon the original paired methods across all metrics, our unpaired mode concurrently achieves quantitative and qualitative performance that provides a competitive alternative to the original pairedtrained counterparts. The code and pre-trained models are available at: https://github.com/gosha20777/EGUOT-ISP.git.

ECCV 2026

ViTexQA: A Multi-Frame Temporal Perception Dataset for Video Text Question Answering

Zhentao Guo, Chen Duan, Tongkun Guan, Zining Wang, Kai Zhou, Pengfei Yan

Despite remarkable progress in multimodal understanding,current MLLMs still exhibit limitations in video text understanding, par-ticularly when semantics emerge through the integration of temporallydistributed textual cues across multiple frames. This perception challengefundamentally differs from static image text understanding, yet existingdatasets fail to capture: the vast majority of questions remain answerablefrom single frames, inadequately reflecting real-world video text compre-hension demands. To address this, we present ViTexQA, a large-scalevideo-text QA dataset, and FrameThinker for robust multi-frame tem-poral reasoning. We build ViTexQA via a quality-controlled Chain-of-Thought (CoT) annotation pipeline boosted with temporal constraints;all its QA pairs demand cross-frame text fusion to solve, enforcing truetemporal reliance. FrameThinker adopts two-stage training for explicittemporal modeling: CoT-Guided Supervised Fine-Tuning (SFT) gen-erates frame-aware reasoning chains, followed by Temporally-groundedReinforcement Learning (RL) optimized with multi-frame coherence re-wards. Evaluations show our method outperforms SOTA baselines onViTexQA, lifting ROUGE-L by 6.3%.

ECCV 2026

Evidence-Backed Video Question Answering

Shijie Wang, Honglu Zhou, Ziyang Wang, Ran Xu, Caiming Xiong, Silvio Savarese, Chen Sun, Juan Carlos Niebles

Current Video Large Language Models (Video LLMs) excel in question answering (QA) but largely operate as black boxes, providing textual answers without verifiable visual grounding. Existing explainability efforts rely on textual rationales or sparse bounding boxes, which struggle to capture complex video dynamics such as occlusions and nonrigid deformations. We propose Evidence-Backed Video Question Answering (E-VQA), a novel task requiring models to jointly output a semantic answer and precise spatio-temporal evidence: temporal segments and dense, tracked object segmentation masklets. To support this, we introduce ST-Evidence, the first human-verified benchmark for both discriminative and generative pixel-level grounding. Evaluations of state-of-the-art models reveal a critical decoupling between QA accuracy and true visual perception that scaling alone fails to bridge. To address this, we develop scalable, automated generation pipelines to create ST-Evidence-Instruct, a 160k-scale dataset bridging highlevel reasoning with fine-grained grounding. Fine-tuning grounded Video LLMs on this data yields substantial gains over the corresponding sizematched UniPixel baselines (e.g., +27.2 t-mean and +13.8 J&F on a 7B model), establishing a robust baseline for explainable, evidencebacked video understanding. Code and data are available at https: //github.com/SalesforceAIResearch/EVQA.

ECCV 2026

SHReg: Strictly Rotation-Equivariant Point Cloud Registration via Spherical Harmonics

Chongjian Wang, Junjie Gao

Point cloud registration critically depends on local featuresthat are both distinctive and robust to arbitrary 3D rotations. Existinglearning-based methods typically approximate rotation invariance viafragile local reference frames or extensive data augmentation, providingonly empirical invariance and often degrading under unseen rotationaltransformations. In this paper, we propose SHReg, a strictly rotation-equivariant point cloud registration framework grounded in the repre-sentation theory of SO(3). By representing local geometric features asirreducible representations of SO(3), SHReg guarantees exact equivari-ance under arbitrary rotations without relying on local reference frames.Built upon a spherical-harmonics-based equivariant backbone, SHRegjointly learns rotation-invariant descriptors for robust correspondencematching and rotation-equivariant features that preserve fine-grained ori-entation information. The preserved equivariant structure enables eachcorrespondence to directly hypothesize a rigid transformation, reducingreliance on large-scale hypothesis sampling in conventional RANSAC-based pipelines and leading to improved robustness under challengingrotational variations. Extensive experiments on 3DMatch, 3DLoMatch,and KITTI demonstrate that SHReg consistently outperforms state-of-the-art methods in registration accuracy, particularly under large rota-tional perturbations.

ECCV 2026

Beyond Disjoint Tasks: Towards More Natural Continual Learning for Vision-Language Models

Xiang Xu, Yiyang Su, Tianchen Zhao, Zheng Zhang, Zhuowen Tu, Anil Jain, Jonathan Wu

Continual learning methods for vision-language models are developed on benchmarks where each new task introduces entirely new domain knowledge. Real-world task sequences are more natural: they routinely share visual concepts, language patterns, and even training samples across stages. However, existing mixture-of-expert methods that assign one expert per task with fixed routing can split similar inputs across different experts and degrade performance. We introduce Semantic Overlap-aware Continual Learning (SOCiaL), a simple framework designed to identify and leverage shared structure across tasks. SOCiaL employs a Gaussian mixture model to estimate contextual similarity, generate synthetic replay samples, and guide expert routing. When task contexts are highly similar, their adapters are consolidated and the router is updated accordingly. To study this more realistic continual learning scenario, we also present UCIT-O, a new benchmark with three protocols that progressively increase semantic similarity across tasks. Across both disjoint and overlapping benchmarks, SOCiaL consistently outperforms existing methods, achieving 7.35, 8.04, and 3.07–9.77 points above the strongest baseline on CoIN, UCIT, and UCIT-O while reducing deployed adapters by up to threefold.

ECCV 2026

Can Vision Models Truly Forget? Mirage: Representation-Level Certification of Visual Unlearning

Zhenyu Yu, Yangchen Zeng, Chunlei Meng, Guangzhen Yao, Shuigeng Zhou

Machine unlearning in Vertical Federated Learning (VFL)has attracted growing interest, yet existing methods certify forgettingsolely using output-level metrics. We challenge these works by intro-ducing Mirage, a representation-level auditing framework built fromfour complementary diagnostics. Mirage combines linear probe recovery(LPR), centered kernel alignment (CKA), feature separability scoring,and layer-wise recovery analysis to assess what a representation actuallyretains. Extensive experiments across seven datasets and seven baselinemethods following recent VFL unlearning protocols reveal three key find-ings. (1) Forgetting gap: methods that pass output-level certification stillretain substantial class structure in their representations, with LPR ex-ceeding the retrained baseline by up to 15.4 points. CKA shows thatthese models remain structurally closer to the original than to the re-trained reference, and separability scores indicate persistent geometricdiscrimination. (2) Unlearning trilemma: no existing method simultane-ously achieves high utility, output-level forgetting, and representation-level forgetting. (3) Class-sample asymmetry: class-level forgetting leavesstrong representational traces (LPR exceeding 96% on several datasets),whereas sample-level forgetting is indistinguishable from chance (LPR≈ 50%). Layer-wise analysis further shows that residual class informationpersists across network depths. These findings call for representation-aware evaluation standards in federated unlearning research. Code ispublicly available at https://github.com/YuZhenyuLindy/Mirage.

ECCV 2026

Learning Structured Visual Compositional Representations for Weakly Supervised Referring Expression Comprehension

Lian Xu, Mohammed Bennamoun, Farid Boussaid, Hamid Laga, Yulan Guo, Dan Xu

Referring expression comprehension (REC) aims to localize the object in an image described by natural language. In Weakly supervised REC (WREC), existing approaches primarily operate on anchorlevel visual representations. Even when enriched with auxiliary cues, relational interactions remain implicitly encoded within individual anchor features. The resulting visual representation remains flat and unary-only, limiting its ability to align with the structured nature of language. In this work, we propose a Structured Visual Compositional Representation (SVCR) learning framework for WREC. Rather than implicitly encoding relations within unary anchors, the proposed SVCR explicitly models both unary object embeddings and pairwise relational embeddings, forming a structured visual representation space. We further introduce a compositional alignment mechanism that matches unary and pairwise visual representations with their corresponding textual embeddings in a unified manner, enabling compositional visual–textual matching under weak supervision. Extensive experiments on RefCOCO, RefCOCO+, and RefCOCOg show that the proposed SVCR achieves state-of-the-art performance. These results demonstrate the effectiveness of explicit structured visual representations and visual-textual alignment for WREC.1

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.