Every paper checked — nothing hidden

All papers

Every paper checked for brief #4 of 12–19 Oct 2026 — nothing is hidden. Brief papers first.

All17,313 In brief24 More (on-topic)3,333 Not relevant13,754 Not assessed87 Older352 Could not confirm93 Blogs with no paper266 All blogs284

Paper	Topic	Relevance
GRID-Planner: Graph-Structured Relational Reward Shaping for Diffusion Planning The paper introduces a graph-structured relational reward shaping framework for diffusion-based trajectory planning in autonomous driving. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on autonomous driving motion planning, not industrial inspection.
VesperMUSE: Multi-Modal Fusion Vision Transformer for Image Semantic Segmentation Under Diverse Light Conditions The paper presents a multimodal vision transformer framework combining thermal guidance and RGB features for image segmentation in low-light environments. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Fuses RGB, thermal, and surface normals for segmentation without language.
VesperMUSE: Multi-Modal Fusion Vision Transformer for Image Semantic Segmentation Under Diverse Light Conditions The paper presents a multimodal vision transformer framework combining thermal guidance and RGB features for image segmentation in low-light environments. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Addresses autonomous navigation segmentation, not factory or defect inspection.
Data-Efficient ConvNeXt Pretraining for High-Resolution Sentinel-1 GRD SAR Representation Learning The paper introduces a sparse convolutional masked autoencoder framework to improve data-efficient pretraining on radar satellite imagery. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses purely on radar remote sensing imagery, not multimodal models.
Data-Efficient ConvNeXt Pretraining for High-Resolution Sentinel-1 GRD SAR Representation Learning The paper introduces a sparse convolutional masked autoencoder framework to improve data-efficient pretraining on radar satellite imagery. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Covers satellite radar pretraining, not manufacturing visual inspection.
Visual Sanad: Auditing Vision-Language Models for False Acceptance of Qur'anic Citations The paper introduces a benchmark to measure how often vision-language models accept incorrect Qur'anic citations across various visual presentations. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies religious text attribution, not industrial inspection or defect detection.
When Models Misread Ink: Auditing VLMs for Arabic Calligraphy Recognition The paper evaluates vision-language models on Arabic calligraphy recognition and identifies frequent default-class prediction failures. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Evaluates calligraphy recognition, not manufacturing or defect inspection.
FeaTok: Feature-Loss Tokenization for Image Generation The paper trains discrete image tokenizers by reconstructing pretrained self-supervised feature embeddings rather than raw pixel values. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Proposes an image generation tokenizer, not industrial inspection.
FeaTok: Feature-Loss Tokenization for Image Generation The paper trains discrete image tokenizers by reconstructing pretrained self-supervised feature embeddings rather than raw pixel values. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on unimodal image generation rather than cross-modal reasoning.
Retrofitting Recurrent Vision-Language Models from Pretrained Backbones The paper reduces vision-language model parameters by pruning layers and applying depth recurrence over remaining language backbone blocks. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on general vision-language compression, not industrial defect detection.
Pathology foundation models for training-free case retrieval in salivary gland tumors The paper tests whether pathology foundation models can retrieve relevant historical cases for rare salivary gland tumors without task-specific fine-tuning. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on vision-based image retrieval in pathology without cross-modal language integration.
Pathology foundation models for training-free case retrieval in salivary gland tumors The paper tests whether pathology foundation models can retrieve relevant historical cases for rare salivary gland tumors without task-specific fine-tuning. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on pathology cancer retrieval, not industrial quality control.
Beyond the Dominant Label: Recovering Coexisting Lesion Evidence with Multiple Instance Learning The paper introduces a multiple instance learning method to recover coexisting secondary lesions from whole-slide pathology images labeled only by dominant severity. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Develops a multiple instance learning algorithm, not an LLM benchmark.
Beyond the Dominant Label: Recovering Coexisting Lesion Evidence with Multiple Instance Learning The paper introduces a multiple instance learning method to recover coexisting secondary lesions from whole-slide pathology images labeled only by dominant severity. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Does not focus on model compression, quantization, or efficient serving.
Beyond the Dominant Label: Recovering Coexisting Lesion Evidence with Multiple Instance Learning The paper introduces a multiple instance learning method to recover coexisting secondary lesions from whole-slide pathology images labeled only by dominant severity. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on medical histopathology slides, not industrial defect detection.
Beyond the Dominant Label: Recovering Coexisting Lesion Evidence with Multiple Instance Learning The paper introduces a multiple instance learning method to recover coexisting secondary lesions from whole-slide pathology images labeled only by dominant severity. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Does not address LLM safety, prompt injection, or deployment guardrails.
Beyond the Dominant Label: Recovering Coexisting Lesion Evidence with Multiple Instance Learning The paper introduces a multiple instance learning method to recover coexisting secondary lesions from whole-slide pathology images labeled only by dominant severity. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on whole-slide image classification without multimodal text or language components.
SCCL-DT: Semantic-Conflict-Aware Contrastive Learning with Dual-Rule Teacher Distillation for Zero-Shot Chest X-Ray Diagnosis The paper presents a contrastive learning framework that handles false-negative pairs during zero-shot chest X-ray vision-language pre-training. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Applies to medical chest X-ray diagnosis, not manufacturing quality control.
FALCON: FPGA-Accelerated Latency-Certifiable Occupancy Networks for Camera-Based Autonomous Driving The paper develops an FPGA-based vision system that enforces explicit latency bounds for autonomous vehicle 3D occupancy prediction. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on autonomous driving camera networks, excluded from industrial vision.
FALCON: FPGA-Accelerated Latency-Certifiable Occupancy Networks for Camera-Based Autonomous Driving The paper develops an FPGA-based vision system that enforces explicit latency bounds for autonomous vehicle 3D occupancy prediction. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Covers FPGA hardware acceleration for 3D occupancy without vision-language modeling.
REKI: Registering Knowledge in VLAs The paper introduces a compact vision-language-action model using register tokens as persistent memory to reduce control oscillations in robotics. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Does not focus on LLM safety, guardrails, or red-teaming deployment risks.
REKI: Registering Knowledge in VLAs The paper introduces a compact vision-language-action model using register tokens as persistent memory to reduce control oscillations in robotics. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on robotics manipulation, not manufacturing visual inspection.
REKI: Registering Knowledge in VLAs The paper introduces a compact vision-language-action model using register tokens as persistent memory to reduce control oscillations in robotics. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Introduces a robotic policy architecture, not an LLM benchmark.
LAMP: Layerwise Adaptive Mixing of Predictors for Training-Free Few-Shot Vision-Language Adaptation The paper proposes a training-free method combining global and regional representations across encoder layers for few-shot vision-language adaptation. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies few-shot adaptation methods, not industrial visual inspection.
OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use The paper introduces an open-source framework and codebase for building generalist computer-use agents that operate across visual and code interfaces. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Develops computer-use desktop agents, not manufacturing defect detection.
FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting The paper presents a refiner that models motion proposals as heterogeneous temporal graphs to improve trajectory forecasting for autonomous vehicles. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on autonomous driving motion prediction, explicitly excluded from industrial vision.
FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting The paper presents a refiner that models motion proposals as heterogeneous temporal graphs to improve trajectory forecasting for autonomous vehicles. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on trajectory prediction models, not LLM evaluation frameworks.
FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting The paper presents a refiner that models motion proposals as heterogeneous temporal graphs to improve trajectory forecasting for autonomous vehicles. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Does not focus on model compression, quantization, or efficient inference architectures.
FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting The paper presents a refiner that models motion proposals as heterogeneous temporal graphs to improve trajectory forecasting for autonomous vehicles. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on trajectory refinement graphs, not vision-language or multimodal models.
FutureLift: Lifting Trajectory Proposals into Bidirectional Future Graphs for Motion Forecasting The paper presents a refiner that models motion proposals as heterogeneous temporal graphs to improve trajectory forecasting for autonomous vehicles. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Covers autonomous vehicle trajectory forecasting, not LLM deployment safety.
How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure The paper investigates learning physical damping properties from impact audio and distills this knowledge into vision models. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Addresses cross-modal distillation for physics probing, not model compression or serving.
How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure The paper investigates learning physical damping properties from impact audio and distills this knowledge into vision models. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Does not address deployment safety, hallucination, red teaming, or guardrails.
How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure The paper investigates learning physical damping properties from impact audio and distills this knowledge into vision models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Probes physical representations across audio-vision, not general LLM benchmarking.
How Much Physics Can Vision Learn from Audio? Probing and Distilling Damping Structure The paper investigates learning physical damping properties from impact audio and distills this knowledge into vision models. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies physical damping from audio impacts, not production quality control.
WaveLLM: Wavelet-Domain LLM Entropy Modeling for Lossless Image Compression The paper uses wavelet subband decomposition with language models to improve entropy modeling for lossless image compression. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Uses language models as entropy models for image compression rather than multimodal reasoning.
WaveLLM: Wavelet-Domain LLM Entropy Modeling for Lossless Image Compression The paper uses wavelet subband decomposition with language models to improve entropy modeling for lossless image compression. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on lossless image compression, not manufacturing defect detection.
MMSearch-OPD: Easy Made Hard, Hard Made Learnable for Long-Horizon Multimodal Search Agents The paper proposes a training method combining data refinement with reinforcement learning and distillation to improve long-horizon multimodal search agents. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Investigates multimodal search agents, not industrial defect detection.
Denoising Textures is a Different Ballgame Altogether! Analyzes image denoisers on repetitive textures and finds fine-tuning on textures leads to catastrophic forgetting of generic images. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Deals with low-level image denoising without cross-modal or multimodal integration.
Denoising Textures is a Different Ballgame Altogether! Analyzes image denoisers on repetitive textures and finds fine-tuning on textures leads to catastrophic forgetting of generic images. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies generic texture denoising, not industrial inspection or defect detection systems.
FPG-Seg: Annotation-free Retinal Vessel Topology Segmentation Presents an annotation-free retinal vessel segmentation pipeline combining classical filtering, Bayesian deep learning, and graph reconstruction. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Covers medical image segmentation without multimodal text, audio, or cross-modal reasoning.
FPG-Seg: Annotation-free Retinal Vessel Topology Segmentation Presents an annotation-free retinal vessel segmentation pipeline combining classical filtering, Bayesian deep learning, and graph reconstruction. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Applies to retinal vessel segmentation, not manufacturing or industrial inspection.
SCOPE: Semantic Bounded Credit Optimization for Prompt Engineering in Text-to-image Generation Introduces a reinforcement learning objective that groups prompt tokens into semantic spans to refine text-to-image prompt rewriting. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Addresses prompt engineering for text-to-image models, not industrial inspection.
Beyond Pairwise Repulsion: Collective Diversity via Determinant Maximization in Diffusion Sampling Maximizes batch Gram matrix determinants during initial diffusion noise optimization to improve overall sample diversity in image generation. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on text-to-image diffusion diversity, not industrial vision or defect detection.
Beyond Pairwise Repulsion: Collective Diversity via Determinant Maximization in Diffusion Sampling Maximizes batch Gram matrix determinants during initial diffusion noise optimization to improve overall sample diversity in image generation. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on sampling algorithms for image generators rather than multimodal understanding.
Unsupervised Clustering of Public Transport Vehicles Using Deep Morphological Features and Color Histograms Combines vision transformer features with color histograms to cluster public transit vehicles in surveillance video without manual labels. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Uses vision-only features for clustering without multimodal language or cross-modal integration.
Unsupervised Clustering of Public Transport Vehicles Using Deep Morphological Features and Color Histograms Combines vision transformer features with color histograms to cluster public transit vehicles in surveillance video without manual labels. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Examines urban bus surveillance clustering, not industrial manufacturing quality control.
MJev-Spatial: System One Spatial Intelligence via Native Attention Matching Frames spatial understanding tasks as candidate classification using pretrained attention matching to enable low-latency spatial decisions. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies spatial intelligence classification, not industrial defect detection.
Routing Policy Optimization for Online Diffusion Reinforcement Learning Proposes an online reinforcement learning approach that uses attention routing as stochastic exploration to align visual diffusion generators with rewards. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Studies diffusion policy optimization, not manufacturing inspection or quality control.
Routing Policy Optimization for Online Diffusion Reinforcement Learning Proposes an online reinforcement learning approach that uses attention routing as stochastic exploration to align visual diffusion generators with rewards. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Focuses on diffusion exploration algorithms rather than multimodal reasoning or representation.
Grounded Human-Attributed Description and Activity Recognition in Videos Introduces a benchmark and method to detect, describe, and recognize activities for every person visible in complex multi-person video scenes. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Examines human activity recognition, not industrial inspection or defect detection.
ARC-SORT: Adaptive, Robust, Compensated Multi-Object Tracking with Interacting Multiple Models Presents a real-time tracking-by-detection method that dynamically scales search regions based on predictive motion uncertainty. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on sports broadcast tracking, not industrial inspection or defect detection.
ARC-SORT: Adaptive, Robust, Compensated Multi-Object Tracking with Interacting Multiple Models Presents a real-time tracking-by-detection method that dynamically scales search regions based on predictive motion uncertainty. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Addresses multi-object tracking via Kalman filters without multimodal text-image modeling.
Velox: Spatial Memory Makes World Models Efficient Introduces a geometry-based attention router that selects visual memory contexts to speed up chunked video world model generation. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Develops video world models, not industrial defect detection.
Not Just a Subject: Capturing Adaptive Scene-Specific Variance for Visual Storytelling Presents a training-free framework that balances identity consistency and narrative scene adaptation for multi-frame visual storytelling. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on narrative visual storytelling, not industrial defect detection.
Dense4D: Pixel-Level 4D Human Correspondence Introduces a feed-forward model that predicts dense 4D point correspondences across articulated human bodies from monocular RGB images. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Focuses on 4D human body correspondence, not industrial inspection.
Dense4D: Pixel-Level 4D Human Correspondence Introduces a feed-forward model that predicts dense 4D point correspondences across articulated human bodies from monocular RGB images. Preprint	Multimodal models via arxiv · 17 Oct, 06:18	Not relevant Covers purely monocular geometric point correspondence, not multimodal integration.
Speculative Forcing for Efficient Autoregressive Video Generation Accelerates autoregressive video generation by verifying and freezing intermediate tokens early to reduce redundant computation during denoising. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Addresses video generation efficiency, not industrial defect detection.
SUPRA-TEX: 3D Generation with Super-Aligned Texture Proposes a framework to dress 3D meshes with aligned materials by decoupling visible surface estimation from observation-guided 3D completion. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Proposes 3D mesh texturing techniques, not manufacturing quality control.
DK-VLM: Geometry as Position for Spatial Reasoning in Vision-Language Models Presents a dual-kernel attention mechanism that uses 3D geometric positions to improve spatial reasoning in vision-language models. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Addresses generic 3D spatial reasoning in VLMs, not industrial inspection.
Grounding Generative Verification in 3D Content Introduces a benchmark spanning ten tasks to evaluate multimodal model verification and rejection explanations on generated 3D content. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:18	Not relevant Evaluates 3D generation verification, not industrial visual defect detection.

4,261–4,320 of 17,313← Previous Next →