ECCV2026 Object Detection AI paper notes paper summaries Anomaly Detection Few-/Zero-Shot Learning Multimodal/VLM Adversarial Robustness Alignment/RLHF

🎯 Object Detection¶

🎞️ ECCV2026 · 79 paper notes

📌 Same area in other venues: 🧠 NeurIPS2026 (1) · 📷 CVPR2026 (99) · 🔬 ICLR2026 (30) · 🧪 ICML2026 (7) · 🤖 AAAI2026 (29) · 🧠 NeurIPS2025 (27)

🔥 Top topics: Object Detection ×25 · Anomaly Detection ×14 · Few-/Zero-Shot Learning ×13 · Multimodal/VLM ×5 · Adversarial Robustness ×4

RT-DETRv4: Painlessly Furthering Real-Time Object Detection with Vision Foundation Models: RT-DETRv4 supervises deep AIFI features with a frozen vision foundation model during training and adapts supervision strength to the module's gradient share, raising the L variant from DEIM-L's 54.7 to 55.4 COCO AP while retaining 8.07 ms inference latency.
M4-SAR: A Multi-Resolution, Multi-Polarization, Multi-Scene, Multi-Source Dataset and Benchmark for optical-SAR Object Detection: This paper proposes M4-SAR—the first large-scale optical-SAR paired oriented object detection dataset (112,174 image pairs, 981,862 instances, 6 categories), accompanied by a unified evaluation tool MSRODet and an end-to-end fusion detection framework E2E-OSDet (FAM + CMIM + AFM). After fusion, mAP improves by up to 5.1% compared to the best single-source modality, with particularly significant gains in complex environments.
Low-latency Event-based Object Detection with Spatially-Sparse Linear Attention: SSLA decomposes the global state of linear attention into geometrically routed local substates, enabling parallel training and event-by-event recurrent inference for low-latency detection; SSLA-L achieves 0.375 mAP on Gen1 with 0.724 MFLOPS per event and a measured single-core processing latency of 7.20 microseconds.
VLOD-TTA: Test-Time Adaptation of Vision-Language Object Detectors: To address the performance degradation of vision-language object detectors (such as YOLO-World and Grounding DINO) under test-time distribution shifts, VLOD-TTA utilizes "IoU-weighted entropy minimization" to focus adaptation on spatially dense clusters of overlapping bounding box proposals, and "image-conditioned prompt selection" to select the text prompts that best match the current image. By updating only a minimal number of adapter parameters, it consistently outperforms existing TTA baselines and previous specialized methods under diverse shifts—including artistic styles, adverse driving conditions, low-light, and common degradations—at a much lower computational cost.
DroneFINE: Domain-Aware Parameter-Efficient Fine-Tuning of Vision-Language Detectors for Drone Images: Building on a frozen GroundingDINO with only 1.54M trainable parameters (5.6% of the 27.5M used by full fine-tuning), DroneFINE combines two domain-aware modules — the foreground-aware dynamic multi-path convolution adapter HyperAdapter and the background-vocabulary gated query re-ranking module SemanticGate — and reaches 38.4 mAP / 61.2 mAP50 on VisDrone and 26.5 mAP / 43.6 mAP50 on UAVDT, matching (and on UAVDT surpassing) full fine-tuning while preserving generalization on COCO.
RGBT-GroundBench: Visual Grounding Beyond RGB in Complex Real-World Scenarios: The paper builds RGBT-GroundBench, the first large-scale RGB-Thermal visual grounding benchmark for complex real-world scenarios (21,535 RGB-TIR pairs, 38,760 referred instances, three-level annotations), defines a unified protocol supporting RGB-only, TIR-only, and RGB+TIR inputs, benchmarks 11 VG models under it, and proposes the reference baseline RGBT-VGNet, which attains the best [email protected] on all three sub-datasets.
What is the Right Embedding Space for Contrastive Learning in Referring Expression Counting?: The paper moves contrastive learning for Referring Expression Counting (REC) out of the image-text embedding space and entirely into the image embedding space, using bipartite matching to define positive and negative sets on the fly, expanding the negative pool from a handful of referring expressions to hundreds of in-image visual tokens, and using only positive-class tokens as anchors; the resulting plug-in module, C-REX, improves three detection-based REC architectures consistently, cutting MAE by up to 28% and RMSE by up to 24.5% on REC-8K, and also transfers to class-agnostic counting.
DetPO: In-Context Learning with Multi-Modal LLMs for Few-Shot Object Detection: DetPO converts a few annotated images into repeatedly validated textual class definitions, using false positives to narrow concepts and false negatives to broaden them, improving few-shot detection with frozen multimodal large language models and optionally refining detection rankings through per-box VQA re-scoring.
MOCHA: Multi-modal Objects-aware Cross-arcHitecture Alignment: MOCHA has a frozen LLaVa teacher produce object-level multimodal embeddings conditioned on "this box plus this class name," maps the multi-scale region features of a YOLO student into that space with a channel-attention translation module, and supervises the result with a dual objective combining pointwise alignment and relational geometry preservation, lifting a 3.2M-parameter detector by 10.1 points on average over the YOLOv8n baseline across four personalized detection benchmarks at only 3 ms/image of extra inference cost.
Rethinking IRSTD: Single-Point Supervision Guided Encoder-only Framework is Enough for Infrared Small Target Detection: SPIRE replaces contour segmentation with centroid response regression, using single points to construct structured local supervision, a lightweight encoder-only network to predict heatmaps, and direct peak extraction to achieve 97.05% F1 on SIRST-UAVB with approximately 0.29M parameters.
CountEx: Fine-Grained Counting via Exemplars and Exclusion: CountEx jointly encodes what to count and what to exclude as detection queries, retaining shared features while suppressing negative-exclusive patterns to resolve fine-grained ambiguity, and achieves 26.61 and 12.72 MAE in the novel- and known-category CoCount settings.
Domain Adaptive Object Detection via Dual-Stream Bilevel-Cycle Optimization: DSBCO splits cycle self-training (CST) into two cycle streams — one for classification and one for regression — and builds them on top of Mean Teacher, replacing the unstable sample-kernel inversion with a feature-dimension ridge-regression projection that is verified on the source domain, plus a normalization of regression offsets; it lifts the FCOS self-training baseline from 41.2 to 64.7 mAP on Cityscapes→Foggy Cityscapes.
IoUCert: Robustness Verification for Anchor-based Object Detectors: IoUCert casts "the IoU of a single-object detection stays above a threshold" as a verifiable property: a coordinate transformation removes the non-linear box-decoding map from the relaxation chain, and a proof that the IoU extremum can only sit at finitely many critical points turns the IoU bounds into an exact enumeration over at most 169 candidates, which together with optimal LeakyReLU relaxations yields the first formal robustness verification of real anchor-based detectors such as SSD, YOLOv2 and YOLOv3.
REAL-OW: Rehearsal-free Open World Object Detection with Low-Rank Adaptation and Dual-Stage Objectness Modeling: REAL-OW is the first strictly rehearsal-free open-world object detection framework — it stores no historical samples at all: general knowledge goes into General Adapters in the backbone, task-specific knowledge into per-task Specific Adapters in the decoder, and a dual-stage objectness modeling (DSOM) with a dispersion-calibrated CG-NLL distance keeps the known/unknown boundary stable, surpassing every exemplar-replay method on M-OWODB and S-OWODB with only 2.0M trainable parameters.
Rethinking Prototype-based Similarity Learning for Few-Shot Object Detection: Addressing two fundamental bottlenecks of prototype-based similarity learning in few-shot object detection—category confusion caused by inter-class similarity margin collapse, and inaccurate localization due to similarity embeddings lacking spatial geometric cues—this paper proposes Text-Anchored Semantic Masking (TSMa) to suppress style-driven spurious similarities using textual features to enlarge inter-class margins, and Stage-Aligned Hierarchical Autoregressive Regression (SHARe) to achieve progressive bounding box refinement by injecting multi-level visual features in reverse order of ViT depth. The proposed model significantly updates the State-of-the-Art (SOTA) on COCO with a substantial gain of +10.1 nAP.
Robust Onion: Peeling Open Vocab Object Detectors Under Noise: Robust Onion peels open-vocabulary object detectors layer by layer using controlled synthetic degradations (turbulence, motion blur, pixelation), locating the failing layer through cross-severity feature overlap: robustness turns out to be governed almost entirely by vision-backbone architecture and depth (shallow layers fail first, while the enhancer and the language side contribute little). The findings are compiled into NN & TK0, a plug-and-play module with only 2.41M trainable parameters that matches end-to-end fine-tuning on real driving datasets with 96× fewer trainable parameters.
RT-DocLayout: Real-Time End-to-End Document Layout Analysis with Reading Order in the Wild: RT-DocLayout unifies layout element classification, bounding-box regression, pixel-level segmentation, and reading order prediction inside a single query-based decoder built on RT-DETR, reaching the best full-document parsing accuracy on OmniDocBench v1.5 and Real5-OmniDocBench (94.50 / 92.05 Overall) with only 33M parameters, while running at 132.1 FPS on an A100 at batch size 32.
SupGRPO: Enhancing GRPO with Matching-based Online SFT for Text Spotting: This paper jointly fine-tunes a multimodal large language model with GRPO and a matching-based online SFT that acts only on coordinate tokens: the SFT branch supplies the token-level localization supervision GRPO lacks, while the matching step removes the instance-order prior that standard SFT imposes on an unordered set of text instances, improving both detection and recognition on Total-Text, ICDAR 2015, CTW1500 and the newly built artistic-text dataset ATS.
The Label Imitation Game: Turing Test Network for Zero-Shot Pseudo-Label Pruning: This work formulates "pseudo-label pruning" as a Turing-test-style adversarial interrogation. It trains a lightweight "judge network" (Turing Test Network, TTN) trained solely on image classification data, yet capable of zero-shot transfer to evaluate object detection pseudo-labels. By leveraging dataset-level semantic context to identify and discard systematic hallucinations of foundation models, it improves the F1 score of the worst-performing categories by 28% (44% after fine-tuning).
Adaptive Spectrum-Aware Feature Disentangled Network for Small Object Detection: SFDNet explicitly disentangles features into low, middle, and high-frequency components in the spectral domain. It designs specialized Mamba scanning strategies for each spectrum to perform context modeling, followed by adaptive fusion. Meanwhile, it constructs class-level prototypes to compress the feature distribution of intra-class objects via contrastive distillation, significantly outperforming previous SOTA methods on AI-TOD, SODA-D, and SODA-A.
CMDS-AD: Cross-Modal Dual-Stream Decoupling for Few-Shot Anomaly Detection: This paper proposes CMDS-AD, a cross-modal dual-stream decoupling framework for few-shot multimodal anomaly detection. By repurposing a pretrained diffusion normal estimator as a non-linear low-pass filter, a pure low-frequency auxiliary estimation stream is constructed to anchor global structural templates. This allows the real stream, which retains coupled high- and low-frequency components, to more accurately capture local micro-defects. Combined with a Coordinate-Aware Hierarchical Feature Mapper and a multiplicative anomaly scoring mechanism, the method achieves an absolute improvement of 5.7% and 7.7% in I-AUROC under the 1-shot setting on MVTec 3D-AD and EyeCandies, respectively.
Deformable Triangle Splatting: Flexible Primitives for Real-Time Radiance Field Rendering: Addressing the over-tessellation and blurring bottlenecks of rigid or convex primitives on curved and concave boundaries, this paper proposes Deformable Triangle Splatting (DETRIS), which equips triangle primitives with learnable barycentric boundary control points, winding number testing, and polynomial smooth-minimum distance fields to achieve real-time, view-independent non-convex radiance field rendering.
Denoising-Enhanced Coarse-to-Fine Infrared Small Target Detection with Attention Prior-Guided Knowledge Distillation: This paper proposes ECFNet, a coarse-to-fine infrared small target detection framework. In the coarse stage, RBCN reduces the full-image dense prediction to grid binary classification and introduces Denoising-Assisted Training (DAT) to force the network to learn target-background context. In the fine stage, a lightweight detector is utilized with Attention Prior-Guided Knowledge Distillation (APKD) to enable the student model to focus on key target regions, achieving SOTA performance on three infrared datasets with only 31.7G FLOPs.
Detecting Backdoors in Object Detection via Pre-NMS Prediction Distribution Shift: This paper introduces DistScan, which reveals for the first time that backdoor injection systematically distorts an object detector's pre-NMS prediction class distribution on clean inputs away from the training class prior, achieving trigger-free, weight-free, and re-training-free black-box backdoor detection via Jensen-Shannon divergence.
Diagnosing Aerial-View Object Detectors with Foundational Image Generative Models: This paper introduces a synthetic diagnostic framework powered by state-of-the-art foundation image generative models (Imagen 3) that couples an explicit attribute taxonomy, text-guided synthesis, complementary image editing, and human-in-the-loop verification to pinpoint detector vulnerabilities across scenes, enabling targeted supplementation with minimal real data to achieve up to 13.4% AP50 gains.
DualCount: Structurally Consistent Density and Point Modeling for Zero-Shot Object Counting: Addressing spatial ambiguity and background leakage in zero-shot object counting caused by weakly regulated density regression, DualCount presents an instance-aware dual-decoder architecture that softly decomposes the density map using predicted instance points and enforces per-instance mass conservation and center-of-mass alignment, establishing new state-of-the-art results on FSC-147 and PUCPR+.
Efficient RGB-T Object Detection via Sparse Cross-Modality Fusion: A sparse fusion framework, SFEDet, is proposed. It first filters candidate foreground regions across the entire image with high recall using a lightweight single-modality detector equivalent to YOLOv8-Small, and then performs precise verification and multi-step bounding box refinement via cross-modality fusion only on sparse RoIs. This achieves detection accuracy comparable to or even better than the state-of-the-art (SOTA) at approximately 1/3 of the computational cost and 1/5 of the parameter count.
FineEdit: Fine-Grained Image Edit with Bounding Box Guidance: FineEdit employs early channel-wise concatenation and deep side-adapter feature injection to embed bounding-box spatial priors into a diffusion transformer, complemented by a decoupled foreground-background reinforcement learning post-training strategy to deliver precise localized editing without background leakage.
Following the Flow: Advection-Consistent Modeling for Event-based Small Object Detection: PACT explicitly models the evolution of event stream features as a physical transport process constrained by the advection equation. Through trajectory-level consistency, weak target responses accumulate along the velocity field while noise naturally attenuates, improving the IoU from 55.18% to 75.90% on the EV-UAV dataset.
FSDC-DETR: A Frequency-Spatial Domain Collaborative DETR for Small-Object Detection: To tackle high-frequency degradation and spectral aliasing caused by rigid spatial aggregation and downsampling in small object detection, FSDC-DETR introduces an explicit frequency-spatial collaborative framework combining dual-branch adaptive fusion, shunt cross-scale feature interaction, and dynamic wavelet downsampling, establishing new state-of-the-art results of 31.1 AP on VisDrone-DET2019 and 32.3 AP on AITODv2.
Fully Rotation-Equivariant Spectral-Spatial Learning for Multispectral Object Detection: To resolve the bottlenecks of discrete spectral processing, scale-dependent cross-level fusion degradation, and geometric disorientation in multispectral object detection, FressDet introduces a fully rotation-equivariant framework featuring order-preserving continuous spectral implicit warping and dual-prototype consistency weighting, establishing a new SOTA with 93% fewer parameters.
Global Logic and Local Search: Dual-Stream Multimodal In-Context Learning for Verifiable Industrial Anomaly Detection: To tackle defect sample scarcity, visual hallucinations, and input resolution bottlenecks in cold-start industrial anomaly detection, GLLS introduces a training-free dual-stream multimodal in-context verification framework that unifies offline part-aware atlas construction with online SAM 3 global structural audits and MCTS-budgeted local evidence search, making diagnostic verdicts explicitly verifiable and traceable to pixel evidence.
Is Monitoring Enough? Strategic Agent Selection For Stealthy Attack in Multi-Agent Discussions: Under a novel discussion-monitored scenario where anomaly detectors continuously inspect inter-agent communications and block overt adversarial messages, this paper proposes an adversarial-aware extension of the Friedkin-Johnsen opinion dynamics model and a Stackelberg game formulation that strategically selects hijacked and target agents to induce maximal consensus deviation under negligible detection rates.
LogiCo: A Unified Framework for Logical and Structural Anomaly Detection: LogiCo proposes a unified component-level feature reconstruction paradigm. By discretizing pre-trained features into a component-level space and synthesizing pseudo-logical anomalies to train a reconstruction network, while introducing cross-attention structural reconstruction and a segmentation map discriminator, it simultaneously detects logical and structural anomalies while maintaining spatial structures. It achieves SOTA on four benchmarks, including MVTec-LOCO.
MATCH: Flow Matching for Multi-View Anomaly Detection: MATCH is the first Flow Matching-based multi-view anomaly detection method. It performs density estimation on pre-trained feature spaces using continuous normalizing flows in the form of ODEs. By omitting the divergence term, it achieves real-time inference (18.77 FPS) and obtains SOTA detection and segmentation performance on both the Real-IAD and MANTA-Tiny benchmarks.
Orthogonal Knowledge Refreshing for Domain-Incremental Object Detection: To tackle catastrophic forgetting and inter-domain interference in exemplar-free domain-incremental object detection, this paper proposes Orthogonal Knowledge Refreshing (OKR), which incrementally constructs isolated low-rank branches for new domains, projects gradient updates onto the orthogonal complement of historical subspaces, and enforces topology-aware consistency to align cross-domain semantic prototypes without any domain routing at inference.
PADFormer: Pose-agnostic Anomaly Detection from Sparse View Images: Tackling the challenge of inspecting objects under arbitrary viewpoints with only sparse unposed reference images, PADFormer bypasses cumbersome 3D reconstruction and introduces a 2D ViT-based cross-view masked reconstruction framework that integrates spatial feature alignment, dynamic patch selection, and anomaly-agnostic prior tokens to detect defects via multi-pass ensemble inference.
Proximity-CLIP: Text-Guided Semantic Proximity Learning for Zero-Shot Anomaly Detection: Addressing the geometric dilemma where enforcing strict orthogonality between normal and anomalous text prototypes disrupts pre-trained continuous metric structures, Proximity-CLIP visually calibrates dynamic semantic margins and deploys an Anomaly Query Module to actively retrieve subtle local defects in zero-shot anomaly detection.
PS-MOT: Cultivating Instance Awareness from Point Seeds for Multi-Object Tracking: This work replaces dense per-frame bounding box annotations in multi-object tracking with "one point per object" point annotations. Through a three-layer design comprising SAM-in-the-loop, frequency-domain boundary hallucination, and uncertainty Gaussian loss, scale-less point seeds are progressively "cultivated" into scale-accurate, identity-continuous instance representations. Under pure point supervision, it achieves 52.3 HOTA on DanceTrack, reducing annotation costs by approximately 64%.
ReFP-AD: Rectified Flow Preconditioning for Energy-Based Anomaly Detection: Addressing the core bottleneck where geometric anisotropy in high-dimensional foundation model token spaces destabilizes finite-step MCMC sampling in Energy-Based Models (EBMs), ReFP-AD introduces an optimal transport rectified flow preconditioning mechanism that maps uncompressed 1536-D features into an isotropic latent space, enabling stable preconditioned SGLD training and setting a new state-of-the-art for unified EBM anomaly detection.
Rethinking Continual Anomaly Detection on the Edge: Benchmarking Under Realistic Industrial Conditions: This paper constructs the first unified benchmark to perform head-to-head comparisons, continuous drift evaluation, and edge-side efficiency tests on existing continual anomaly detection (CAD) methods. The evaluation reveals that dedicated CAD methods do not outperform traditional anomaly detection baselines equipped with simple experience replay. Based on this finding, the paper proposes DINOSaur—a training-free method featuring a fully frozen DINOv3 backbone and a spatially indexed coreset memory bank, achieving state-of-the-art (SOTA) performance under zero-forgetting constraints with sub-100ms edge inference latency.
ReynoldsFlow: Physics-Inspired Spatiotemporal Flow Representation for Video Understanding: Addressing the brightness constancy bottleneck and hefty computational costs of traditional optical flow and deep spatiotemporal models, ReynoldsFlow introduces an unsupervised, training-free representation based on the Reynolds transport theorem and Helmholtz-Hodge decomposition that decouples motion into curl-free and divergence-free components while fusing motion magnitude with appearance cues, delivering substantial accuracy gains and real-time efficiency across pose estimation, action recognition, and tiny object detection.
RiO-DETR: DETR for Real-time Oriented Object Detection: Addressing high latency, semantic-geometric entanglement, and periodic optimization instability in prior oriented DETRs, RiO-DETR introduces geometry-decoupled query encoding, rotation-rectified orthogonal attention, decoupled periodic refinement, and oriented dense O2O supervision, establishing the first end-to-end real-time oriented detection transformer with 2.7–29.9 ms latency and 78.4%–81.8% AP50 on DOTA-1.0.
Robust Zero-shot Anomaly Detection under Limited Auxiliary Anomaly Priors: DIVE is the first to investigate the scenario of zero-shot anomaly detection with limited anomaly priors in auxiliary data. By employing a shallow-to-deep text embedding injection strategy, it enables the vision encoder to abstract cross-domain general anomaly concepts from limited auxiliary anomaly patterns. It also introduces a decoupling mechanism to eliminate the interference of object semantics on vision embeddings. When using the texture dataset DTD as auxiliary data, DIVE outperforms SOTA baselines on 12 target datasets by up to 28.5% (AP) in classification and 47.0% (AUPRO) in segmentation, while compressing the performance degradation caused by insufficient auxiliary data diversity by approximately 46.5%-77.3%.
RT-SDGOD: Real-Time Single-Domain Generalized Object Detection: Addressing the issue that real-time detectors suffer severe degradation dominated by missed detections under adverse weather and lighting shifts, this paper proposes RT-SDGDet, a zero-inference-overhead framework that expands and stabilizes object-level discriminative evidence via one-to-many query grouping, deformable attention diversity regularization (DEDL), and dual-view consistency alignment (DvECL).
SelfMOTR: Revisiting MOTR with Self-Generating Detection Priors: To tackle query interference between object detection and temporal association in end-to-end transformer trackers, SelfMOTR generates internal 4D anchor priors and confidence-modulated proposal queries via a lightweight detection forward pass, achieving state-of-the-art detector-free tracking performance on DanceTrack and Bird Flock Tracking.
Towards High-Resolution Visual Perception via Hierarchical Entity Exploration: To tackle entity fragmentation and background interference caused by conventional geometric image partitioning in high-resolution visual perception, this paper proposes Hierarchical Entity Exploration (HEE)—a training-free, model-agnostic framework that structures visual search around object-level entities, clusters them into coherent candidate subregions, and leverages dual scoring with confidence-guided backtracking to achieve superior accuracy and efficiency.
Towards Sparsely Annotated Open World Object Detection: To resolve contradictory supervision arising from ambiguous unlabeled regions in real-world scenarios, this paper introduces Sparsely Annotated Open-World Object Detection (SA-OWOD) and proposes Dual-Perspective Object Discovery (DPOD), which leverages known target recovery and cross-view semantic disagreement to significantly boost unknown object recall while maintaining high known detection accuracy.
\(S^{2}\)-FracMix: Label-Preserving Self-Saliency Mixup Augmentation: S²-FracMix proposes a label-preserving self-saliency mixup augmentation framework: extracting multi-scale saliency patches within the same image, then mixing them back into the original image after rotation, blurring, and fractal texture injection, thereby avoiding cross-sample semantic interference; when paired with a multi-mode high-level mixing strategy, it comprehensively outperforms SOTA methods such as AdAutoMix on 7 benchmarks including classification, detection, and robustness, while introducing negligible training overhead.
A Simple Baseline with Placement Prior for Point-Supervised Oriented Object Detection: PlacePriorBase combines neighborhood angle alignment, image-level size regularity, and training-free watershed box targets to build an inexpensive point-supervised detector, while its transferable losses improve stronger detectors without making the lightweight baseline itself the most accurate model.
Back-Tracking from Clarity: Self-Learning to See Text from Afar: The method propagates reliable detections of clear text backward into earlier, degraded video frames, combining point tracking with carrier geometry to train an unchanged single-frame student; its SceneText50 tracking AUC reaches 44.71 versus 38.65 for CoTracker3.
Calibrate Before Adapt: Training-Free Pseudo-Label Calibration for Semi-Supervised Cross-Domain Few-Shot Detection: To tackle severe semantic misclassification and low recall in pseudo-labels generated by open-set detectors for semi-supervised cross-domain few-shot object detection (Semi-CDFSOD), this paper introduces a training-free pseudo-label calibration framework that couples center-weighted prototype semantic calibration with SAM-based visual self-prompting to purify and complement unlabeled target-domain supervision.
Cross-token Guidance Transformer for Weakly Supervised Object Localization: To tackle the severe diffuse activation caused by relying exclusively on location tokens without semantic context, this paper proposes the Cross-token Guidance Transformer (CGTR), which employs an Attention Regulation Module (ARM) and a Filter Regulation Module (FRM) to establish cross-scale semantic guidance between class and location tokens, achieving state-of-the-art localization precision on CUB-200-2011 and ILSVRC.
Decoupling Moment from Event for Video Temporal Grounding: Addressing the forced semantic suppression problem caused by directly adopting rigid 2D object detection one-to-one matching in video temporal grounding, ME-DETR decouples sub-event moment semantic exploration from full-event boundary regression via dynamic query role specialization and dual-head synergistic supervision, delivering a +2.43% test mAP gain on QVHighlights without requiring NMS.
Defect-aware Hybrid Prompt Optimization for Zero-Shot Multi-type Anomaly Detection and Segmentation: To tackle CLIP's failure in discerning fine-grained defect categories and its vulnerability to drifting away from natural-language manifolds during tuning, DAPO couples shared learnable anomaly context tokens with fixed literal defect anchors, enabling zero-shot multi-type anomaly segmentation under frozen vision-language encoders.
DeMuS: Learning Decoupled Matching and Scoring for Batch Zero-Shot Industrial Anomaly Detection: Addressing the realistic Batch Zero-Shot Industrial Anomaly Detection setting with small intra-batch references and severe pose variations, DeMuS explicitly decouples physical part matching ("where-to-match") from anomaly semantic scoring ("how-to-score") using two lightweight adapters, eradicating anomaly-to-anomaly shortcut matching while delivering real-time online inference.
DETR is Secretly a Multispectral Detector: Zero-Parameter Adaptation via Semantic Alignment: Challenging the parameter-expanding convention in multispectral detection, this paper introduces ZPA-MDETR, which preserves the unimodal DETR architecture and weights without adding learnable fusion parameters, achieving state-of-the-art performance via asymmetric query-memory organization and parameter-free spatial-channel semantic alignment.
DG-Force: Disentangling and Gathering Forensic Cues is Needed for Image Manipulation Localization: To prevent increasingly realistic generated textures from submerging critical boundary artifacts, DG-Force explicitly disentangles patch-level and edge-level forensic cues, adaptively aggregates them via per-token gating, and transfers multi-scale evidence for highly robust image manipulation localization.
Explicit Semantic–Spatial Alignment for Open-Vocabulary Object Detection: To tackle the semantic-spatial discrepancy and resulting localization degradation in open-vocabulary object detection, ESSA-OVD builds a dual-stream architecture combining frozen CLIP and DINOv2 backbones, progressively injecting high-frequency spatial cues via a Spatial Adapter and Frequency-Aware Fusion, while preserving zero-shot semantic matching through a Pure Dense routing strategy with only 3M extra trainable parameters.
Fast Dynamic Prototypes for Unsupervised Anomaly Detection and Localization: Addressing the storage bottleneck of matching-based methods and the implicit opacity of reconstruction-based frameworks in industrial unsupervised anomaly detection, this paper introduces Fast Dynamic Prototype (FDP)—an explicit scheme pairing test-conditioned dynamic bank adaptation with dual-backbone artifact suppression and hybrid residual learning to achieve state-of-the-art accuracy at inference speeds up to 477 FPS.
FuDU: A Fuzzy Dual-dimension Uncertainty Framework for Streaming Active Learning in Industrial Defect Detection: Addressing boundary ambiguity and out-of-distribution outlier risks in streaming industrial defect detection, FuDU combines dynamic backbone prototypes with dual-entropy detection head evaluators to quantify global and defect uncertainties, driving expert knowledge-based fuzzy inference to achieve 96.0% mAP and 99.3% recall at minimal annotation cost.
GAP-Track: Bridging the Resolution Gap for Cross-Resolution RGBT Tracking: Addressing performance degradation on edge devices constrained by low compute and narrow bandwidth, GAP-Track combines training-time masked residual reconstruction, polar-geometric sensitivity loss, and hierarchical knowledge distillation to enable high-precision cross-resolution RGBT tracking with zero inference overhead.
HER-Count: Learning Hyper-Exemplar Representation for Generalized Zero-Shot Object Counting: Addressing the issues where patch-based exemplars fail to cover intra-class diversity and accumulate detector errors while text embeddings lack image-specific instance cues, HER-Count leverages a multimodal large language model (MLLM) to synthesize holistic hyper-exemplar representations (HER), enabling high-accuracy, noise-robust zero-shot counting and localization via hierarchical fusion and global discrimination enhancement.
HLRAD: High-dimensional Latent Representation for Unified Anomaly Detection: HLRAD tackles multi-class unified unsupervised anomaly detection by challenging the conventional low-dimensional bottleneck paradigm, introducing a high-dimensional latent reconstruction framework equipped with a frozen Diffusion Transformer semantic branch and selective gated fusion to eliminate compression information loss and block identity shortcuts.
InfraNet: Quality-Aware RGB Guidance for Infrared Object Detection: To tackle negative transfer and noise contamination from degraded RGB images in adverse conditions, InfraNet establishes an IR-centric asymmetric framework where RGB serves as quality-regulated auxiliary supervision during training via QualGate (suppressing RGB noise and compensating IR features), enabling both zero-overhead IR-only (InfraNet-IR) and high-capacity dual-modal (InfraNet-RGB-IR) inference.
InstaPano: Zero-shot Instance Layout Controlled Panorama Generation Via Global Attention Fusion: InstaPano introduces a training-free framework for zero-shot instance layout-controlled panorama generation, leveraging a Sync-Fuse-Dispatch workflow for global attention fusion and a Conditional Position Mask to eliminate object repetition in large bounding boxes.
MSPL: Multi-Step Pseudo-Labeling for Open-Vocabulary Object Detection: MSPL reformulates pseudo-label generation for open-vocabulary object detection from single-step CLIP image-text alignment into an decodable, three-step reasoning pipeline: "localization verification -> category identification -> background grounding". Utilizing SAM category-agnostic segmentation and MLLM zero-shot inference, it generates high-quality pseudo-labels. These are then used during online training via two contrastive learning signals: region-text alignment and contrastive background learning. It achieves new SOTA performance on both OV-COCO and OV-LVIS, with a 9.4-point gain in novel-class AP50.
OpenPanoD: Aligning Multimodal Prompts and Spherical Representations for Open-Vocabulary Panoramic Detection: To overcome semantic ambiguity from contextual co-occurrence in text prompts and severe geometric distortion from equirectangular projections in panoramic scenes, OpenPanoD introduces a multimodal prompt encoder uniting descriptive text with visual exemplars, paired with a geometry-aware spherical Transformer (GeoFormer) built on a subdivided icosahedral grid for zero-shot 360-degree object detection.
Per‑Object IoU Forecasting for Deadline‑Aware Real‑Time Embedded Detection Control: This paper introduces an ultra-lightweight per-object IoU decay model (OLAP) and a closed-loop runtime detector selection controller (OLAP-C) for streaming detection, deriving closed-form per-object deadlines and optimizing priority-weighted Relative Average Precision to achieve a 56% improvement in deadline prediction accuracy and a 39% gain in streaming AP across Argoverse-HD and MOT17.
PhysFlowNet: Learning Canonical Latent Manifolds via Spatio-Spectral Physics Priors for Underwater Object Detection: Addressing the non-linear feature manifold distortion caused by underwater light attenuation and scattering, PhysFlowNet models representation rectification as a physics-guided Riemannian feature retraction, coupling hierarchical spatio-spectral priors, structure-preserving downsampling, and evidential uncertainty calibration to suppress backscatter false alarms and establish new SOTA benchmarks.
Plug-and-Play Attention Linearization for Pretrained Transformers: This paper introduces LATTE, a training-free attention linearization framework that approximates softmax attention via a second-order Taylor expansion around tunable anchor points paired with Joint Post-Squaring Normalization (JPN), linearizing over half of the attention layers while retaining over 95% of original performance and cutting attention GFLOPs by up to 15×.
Rethinking Detection Calibration: A Coordinate Perspective: Addressing the limitation that conventional detection calibration only measures overall box-level IoU or precision while neglecting independent coordinate alignment and deviation directions, this paper proposes ReDC, a post-hoc calibration framework that integrates coordinate-wise alignment ratio (CAR), confidence re-encoding (CR), and directional displacement estimation (DDE) to deliver fine-grained coordinate-level confidence and directional cues while accurately reconstructing box-level IoU.
Rethinking Pseudo-Labels: Multi-Granularity Supervision for Domain Adaptive Object Detection: Addressing the vulnerability of conventional teacher-student domain adaptive detectors that rely exclusively on noisy proposal-level pseudo-boxes, this paper presents Multi-Granularity Pseudo-Labeling (MGPL), which extracts noise-resilient image-level predictions via a dual-competition aggregation mechanism and refines proposal selection through coarse-to-fine collaborative filtering, achieving 51.4% mAP on Pascal VOC→Clipart1k and outperforming the fully supervised Oracle by 6.4%.
SE-DETR: Explicit Semantic Exploration for Generalizability and Distinguishability in Video Temporal Grounding: SE-DETR resolves the limitations of word fragmentation and critical sparse semantic dilution in video temporal grounding by introducing Semantics-Proxied Alignment (SPA) and self-information-based Temporal Sparsity Modulation (TSM), establishing new state-of-the-art results across five public benchmarks.
Slim-DETR: Real-Time Tiny Object Detection with Efficient Interaction and Gaussian Query: To tackle encoder sequence length bottlenecks and severe query-object mismatch in DETR frameworks for tiny targets, Slim-DETR introduces an Efficient Interaction (EI) encoder with gather-and-inject foreground deformable attention alongside a size-decoupled Gaussian Target Guided (GTG) query selection strategy, achieving 33.9% AP on VisDrone with real-time inference on both RTX 3090 (24.4 ms) and Jetson AGX Orin (40 FPS).
SplitHDR: Saturation-Aware HDR Recovery and Denoising for Real-Time Detection: Addressing the stringent latency constraints and abrupt illumination variations of autonomous driving edge platforms, SplitHDR proposes an input-luminance-adaptive dual-mode architecture that decouples short-exposure highlight recovery in saturated regions from lightweight denoising in unsaturated dominant regions, achieving state-of-the-art downstream detection accuracy (38.38% AP / 58.75% AP50) with merely 2.47k parameters and 3.63 GMACs.
SVCBench: A Streaming Video Counting Benchmark for Spatial-Temporal State Maintenance: SVCBench reframes video counting as a minimal probe to diagnose the "spatial-temporal state maintenance" capabilities of video-language models. By utilizing streaming multi-point queries to observe prediction trajectories and designing three complementary metrics, it reveals systematic deficiencies of current models in periodic event counting and continuous identity tracking.
Task-Agnostic Incremental Vision-Language Object Detection via Prompt Augmentation and Distribution-Aware Fusion: To tackle cross-task prompt interference and LoRA parameter conflicts in incremental vision-language object detection without oracle task priors, TADA introduces training-time stochastic prompt augmentation and test-time Mahalanobis-distance distribution-aware fusion, achieving 64.5 mAP on the ODinW-13 benchmark (+32.0 mAP over DitHub).
VarProtoAD: Variational Prototype-Conditioned Prompting for Zero-Shot Anomaly Detection: VarProtoAD discovers hyperspherical normal and abnormal prototype banks in the frozen multi-layer patch feature space of pre-trained VLMs via a variational von Mises–Fisher (vMF) mixture, and steers learnable context tokens using confidence-gated cross-attention alongside implicit deviation prompts to achieve cross-domain zero-shot anomaly detection and localization.