The paper introduces a structured‑prior‑guided diffusion inpainting framework that enhances traffic sign augmentation by incorporating semantic, appearance, and geometric priors through text prompts, vector templates, and ControlNet. It enforces physical consistency with colour and edge‑structure losses, achieving superior reconstruction fidelity and OCR accuracy compared to larger models. The synthetic data generated improves detection performance for rare traffic sign classes by up to 7.4× over a real‑data‑only baseline.
By Luo Li, Chongchong Huang, Jun Jia, Qiang Gao, Xinlong Liu, Gui Yang, Liang Cao
The paper investigates Automatic Target Recognition (ATR) in Synthetic Aperture Sonar (SAS) imagery, comparing modern convolutional neural networks (CNNs) and transformer-based deep neural networks (DNNs). It examines how factors such as network size, architecture, pretraining methods, data augmentation, and regularization influence performance, aiming to identify the highest-performing model and provide a training roadmap for state‑of‑the‑art SAS‑ATR systems.
By C. J. Moore, Alex Hurt, Jordan Malof
The paper presents a multi‑campaign UAV thermal image dataset for inert ordnance screening, comprising 5,855 labeled image pairs collected in Tennessee across diverse terrains and seasons. The authors trained YOLOV11l and RT‑DETR‑R50 models on 33 m and 15 m altitude data, achieving automated candidate detection, and provided practical guidelines for future humanitarian mine action surveys. The dataset and models aim to aid screening and prioritization for follow‑up technical surveys or EOD assessment, not to replace clearance operations.
By Chad Melton, PhD., Annabelle Kelton
Vision‑Language Models (VLMs) are increasingly replacing traditional OCR for document understanding, but this study shows they often rewrite imperfect text into more plausible forms, a flaw that clean‑text OCR benchmarks miss. The authors created FaithC4, a multilingual perturbation benchmark of 1,455 single‑page documents with scramble, random substitution, and visually similar substitution attacks, and evaluated 15 systems across general‑purpose VLMs, OCR‑specialized VLMs, and traditional OCR pipelines. Results reveal that general‑purpose VLMs suffer up to 6.9 WER points under perturbation, OCR‑specialized VLMs 0.1–3.4 points, and traditional OCR less than 0.8 points on English; probing Qwen3‑VL‑4B shows rewriting occurs only when a perturbed word’s final‑layer representation remains close to the original, with short words (4–6 characters) rewritten up to 10% of the time.
whyItMatters":"The findings highlight a critical limitation of VLMs in document transcription, underscoring the need for robust evaluation benchmarks that capture rewriting behavior beyond clean‑text accuracy."
By Gwang Gook Lee, Kenan Emir Ak, Jay Mohta, Yan Xu, Dimitrios Dimitriadis
MUDIDI is a two-stage framework designed to digitize multilingual dictionaries that are currently only available as scanned images. The first stage assesses character recognition and markup preservation, while the second stage segments dictionary entries and maps them into the SIL Multi-Dictionary Formatter schema. The authors also release a dataset of 30 annotated dictionaries and benchmark OCR, LLM, and VLM systems, finding that LLMs generally outperform others and that providing additional context improves digitization quality.
By David Setiawan, Temuulen Khishigsuren, Milind Agarwal, Pagnarith Pit, Aso Mahmudi, Ekaterina Vylomova
The paper introduces the concept of information density to explain category bias in visual object detection. It finds a strong negative correlation between a category’s information density and its detection accuracy, showing that instance count alone does not account for bias. By incorporating information density into three advanced loss functions, the authors demonstrate reduced model bias and improved overall performance on Pascal VOC, COCO‑LT, and LVIS datasets.
By Ziwei Zhao, Yanxi Lu, Yuwei Hu, Shiyang Su, Mingxuan Wang, Chenyue Zhou, Jiayi Chen, Hehan Li, Xiaoshuai Hao, Andi Zhang, Yanbiao Ma
SliceBridge is a new framework for repairing corrupted slice intervals in T1‑weighted MRI by using rectified flow matching conditioned on surrounding intact slices and their relative positions. It enforces through‑plane consistency by coupling slices within the interval through shared noise, flow time, and synchronized sampling, then reinserts the restored interval without altering other slices. In experiments on 9,877 brain MRIs and 581 external subjects, SliceBridge reduced slice‑to‑slice error by 32.9%–41.3% and improved SSIM across interval lengths, while lowering downstream segmentation volume estimation error from 1.95% to 1.05%.
By Jiheng Li, Michael E. Kim, Trent Schwartz, Gaurav Rudravaram, Derek B. Archer, Timothy J. Hohman, the Alzheimer's Disease Neuroimaging Initiative, Lianrui Zuo, Bennett A. Landman
InstEditSeg is a generative framework that treats medical segmentation as an instruction-driven image editing task. Instead of producing binary masks, it renders a color-coded overlay on the original image guided by textual instructions, leveraging latent diffusion models to align with natural image distributions and reduce domain gaps. The method incorporates a DINOv3 visual encoder and a multi-scale feature pyramid fused into the diffusion U‑Net, and uses a dual‑branch classifier‑free guidance strategy to lower inference cost, achieving competitive accuracy on polyp and skin lesion datasets while improving cross‑domain generalization and multi‑lesion segmentation.
By Ziquan Liu, Zhewei Zhu, Xuyang Shi
The paper introduces AURA, an nnU-Net-based asymmetric supervision strategy for segmenting ultra‑low‑field (0.064 T) pediatric brain MRI. It treats high‑field‑derived (HF) and low‑field‑edited (LF) annotations as distinct observations, anchoring training to the HF mask while gating LF contributions through a reliability mechanism. On a 16‑case development split, AURA achieved Dice scores comparable to HF‑only training and improved boundary metrics, demonstrating its potential for ULF MRI segmentation.
By Ha-Hieu Pham, Dang P. M. Cao, Minh Hoang Pham, Khanh Nguyen Vo Ngoc, Thanh-Huy Nguyen, Ulas Bagci, Huy-Hieu Pham
The paper introduces MiRA, a plug‑in framework that reweights framewise attention in Vision Transformer video models to better capture subtle facial dynamics for expression recognition. MiRA computes frame‑level confidence and intra‑frame concentration from self‑attention maps, redistributing attention toward localized facial cues without adding trainable parameters. Two modes—an exact post‑softmax redistribution and a lightweight flashLite pre‑softmax approximation—are proposed, and experiments on facial expression recognition benchmarks show consistent gains over strong ViT baselines.
By Seongro Yoon, Donghyeon Cho, Jinsun Park, Fran\c{c}ois Br\'emond
AlphaRAD introduces a grounded zero‑shot classification framework for chest radiology that leverages structured medical concepts extracted from reports and a novel α‑Corrected Binary Cross‑Entropy loss to reduce in‑batch noise. It also presents FLaS, a lightweight cross‑modal fusion module that factorizes VLPM representations into independent subspaces, improving spatial grounding without adding parameters. The method achieves state‑of‑the‑art performance on 16 classification benchmarks and sets new records on several grounding, phrase‑grounding, and segmentation datasets.
By Jianzhong You, Yuan Gao, Chris McIntosh
MMTryon is a multi‑modal, multi‑reference virtual try‑on framework that generates high‑quality compositional try‑on results using text instructions and multiple garment images. It addresses three overlooked problems: supporting multiple try‑on items, allowing dressing style specification via text, and eliminating reliance on segmentation models by using a parsing‑free garment encoder and a scalable data generation pipeline. Experiments on high‑resolution benchmarks and in‑the‑wild test sets show MMTryon outperforms state‑of‑the‑art methods qualitatively and quantitatively.
By Xujie Zhang, Ente Lin, Michael Kampffmeyer, Zhenyu Xie, Jiang Li, Ting Liu, Xiaochao Qu, Luoqi Liu, Xiaodan Liang
The paper examines how rewriting verbose, narrative-style queries into concise keyword‑grounded forms improves complex object segmentation in 4D Gaussian representations. By applying a training‑free reinterpretation strategy, the authors reduce linguistic noise while preserving essential semantic anchors. Experiments on HyperNeRF and Neu3D show that rewritten queries boost temporal accuracy from 60.92% to 92.21% and vIoU from 20.08% to 76.94%, with ablation studies confirming the benefits of shorter, keyword‑focused queries.
By Thanh-Khoi Nguyen, Thien-Phuc Tran, Minh-Triet Tran
PointGauss is a 3D-native framework that performs semantic parsing and instance segmentation on 3D Gaussian splatting representations by treating Gaussian primitives as unstructured point sets and extracting scale‑invariant geometric features with Point Transformer V3. It introduces an adaptive region‑of‑interest cropping strategy and an instance‑aware distance‑constrained rasterization pipeline to enable scalable, view‑consistent pixel‑level projections. The authors also release SplatSeg‑360, a cross‑scale benchmark with 32 complex scenes and over 6,300 aligned 2D‑3D masks, and show that PointGauss achieves real‑time performance with state‑of‑the‑art 3D‑mIoU (~90%) and 2D‑mIoU (~80%) scores.
By Wentao Sun, Yiping Chen, John S. Zelek, Jonathan Li
The paper proposes a unified forensics framework that extends traditional binary image manipulation detection to a multiclass setting—distinguishing real, fully synthetic, and tampered images. It adds a segmentation branch for pixel‑level localization of tampered regions, achieving higher classification accuracy and IoU scores compared to recent benchmarks. The authors provide the implementation on GitHub for reproducibility.
By Annalisa Gallina, Marco Fiorucci, Marco Brigo, Federica Battisti, Lamberto Ballan
The paper introduces a dual‑backbone architecture for glass segmentation that combines a frozen foundation model with a learned backbone trained on glass‑specific data. By fusing hierarchical multi‑scale features from both backbones, the method produces accurate segmentation masks and achieves state‑of‑the‑art performance on four benchmark datasets. Ablation studies confirm the benefits of the dual‑backbone design and its generalizability across different backbone choices, while also offering competitive inference speeds, especially with lighter backbones.
By Risto Ojala, Tristan Ellison, Mo Chen
The paper investigates how much clinical text influences pixel‑level predictions in multimodal medical image segmentation. It shows that segmentation performance is largely insensitive to the choice of fusion module, but that the impact of text varies across datasets: removing text severely degrades performance on BUSI and BTMRI, while it has only a marginal effect on ISIC and Kvasir‑SEG. Using an Evidence Decoupling Decoder, the authors reveal that text mainly modulates global semantic context rather than spatial localization, and that the specific semantic components driving sensitivity differ by dataset.
By Ziquan Liu, Zhewei Zhu, Xuyang Shi
The paper presents a stereo 4D Radar framework for 3D object detection that uses geometric disparity between left and right radars to estimate absolute velocity and fuse complementary features. It addresses clutter, ghost reflections, and sparse data issues inherent in raw 4D Radar signals. Experiments on an in‑house dataset show significant gains, improving AP 3D by 8.82 points and AP BEV by 9.0 points over mono‑radar baselines.
By Seung-Hyun Song, Dong-Hee Paek, Woong-Chan Byun, Seung-Hyun Kong
T2LSC-Bench is a new benchmark for evaluating localized semantic control in text-to-image generation, consisting of 50 seed subjects and 1,200 prompt cases per model, producing 7,160 images across six models. The benchmark measures Text-at-Anchor Accuracy, Semantic Subject Preservation, Semantic Leakage Rate, and Conditional Semantic Leakage Rate using a dual‑branch protocol that combines OCR‑VLM verification with structured VLM semantic judgments. Results show that while accurate text rendering remains high, semantic leakage can increase dramatically under stress‑test conditions, and anti‑leakage prompting can reduce leakage without harming rendering accuracy.
By Yan Wang, Xinyi Hou, Weiguo Lin, Junjun Si, Siwei Ma
The paper introduces a two-stage framework for point-supervised change detection that leverages SAM2 priors to generate object-aware candidate masks and refines them with a lightweight CNN and uncertainty-aware loss. In the second stage, a teacher‑student self‑training loop with exponential moving average updates continuously improves pseudo‑labels and model performance. Experiments on WHU-CD, LEVIR-CD, and SYSU-CD show the method surpasses prior weakly supervised approaches and competes with fully supervised ones.
By Hailong Ning, Hao Wang, Yimeng Wang, Tao Lei, Renwei Dian, Asoke K. Nandi