The paper introduces SCR-V2A, a video-to-audio generation framework that focuses on Semantic-Critical Regions (SCRs)—localized visual cues directly related to sound events. SCR-V2A first identifies these regions via sound-related visual localization, then encodes them with an adapter-based dual visual encoder to produce a control representation that captures both global context and fine-grained sound evidence. This representation conditions an audio generation model, and experiments on in-domain and out-of-domain benchmarks show improved distribution matching and audio quality with few trainable parameters.
By Yun Fang, Wei Guo, Zhuonan Liang, Heng Wang, Weidong Cai
RAP (Reconstruct A Poster) is a new dataset and benchmark for generating conference posters from research papers, comprising over 25,000 verified paper‑poster pairs from 2022‑2026. Each pair is annotated with a three‑level content taxonomy and converted into styled HTML/CSS markup through an automated derendering pipeline, enabling objective render‑and‑compare evaluation. The corpus supports both inference and training, with a retrieval‑based PosterRAG pipeline and a supervised‑fine‑tuned model, and is publicly available via an online demo.
By Haonan Han, Yukang Lin, Minliang Lin, Haoxian Tan, Jun-Yan He, Rui Yang, Zhi-Qi Cheng, Jie Hu, Guanglu Wan, Xiu Li
Agentic Image Generation in the Wild pairs an image generator with a web‑search agent to retrieve references beyond the generator’s training data. The authors introduce the Savanna benchmark, comprising 500 prompts derived from recent headlines that are difficult to resolve due to limited visual evidence and confusing search results. They propose the Locate‑Corroborate‑Isolate paradigm and train Ranger‑8B, which outperforms existing frameworks on both search accuracy and generation quality.
By Yifan Li, Bowen Qin, Kun Zhou, Haochen Zhao, Zikang Liu, Zheng Li, Minsheng Hao, Yuhao Cheng, Haosheng Zou, Jeff Chen, Wayne Xin Zhao
MOSS-Image is an open-weight image generation system that combines a 5B single-stream diffusion transformer with an intention-aware prompt enhancement module to produce high-fidelity, text-rich images with precise layout control. The system supports bilingual long-form text rendering and introduces evaluation benchmarks LongLongTextBench and RealLayoutArena. It achieves state‑of‑the‑art performance on LLTB and ranks highly on RLA, with plans to release the model, training recipe, and evaluation framework for the community.
By MOSI. AI
FloodVLM-Pix is a new language‑guided pixel‑level flood understanding model introduced in the arXiv paper "FloodVLM-Pix: Language‑Guided Pixel‑Level Flood Understanding." The authors created the first large‑scale remote sensing flood multi‑modal benchmark, comprising 131,200 image patches and 443,100 textual descriptions of flood severity and evolution. FloodVLM‑Pix incorporates a flood prior module to mitigate attentional bias and a multi‑temporal flood pixel decoder with a general‑specific feature fusion mechanism to improve flood segmentation accuracy.
"whyItMatters":"The paper presents a benchmark and model that enhance both flood description quality and pixel‑level segmentation, addressing shortcomings of existing vision‑language models on flood‑related tasks."
By Chengyang Zhang, Zhihao Wang, Zhiyu Long, Xueqian Wang, Quanming Yao, Wei Wu, Gang Li
ShadowCast is a novel on-the-fly, context-aware shadow augmentation method for remote sensing image segmentation. It synthesizes shadows by extracting masks of elevated objects from segmentation labels and projecting synthetic shadows along an inferred cast-shadow direction, without requiring additional networks or metadata. Experiments on Vaihingen, Potsdam, and LoveDA demonstrate consistent performance gains across various state-of-the-art segmentation architectures and when combined with standard augmentation and regularization techniques.
By Yechan Kim, Dongho Yoon, Jongmin Joo, JongHyun Park, Moongu Jeon
The paper proposes a new evaluation framework for generative image models that uses a generator‑specific classifier trained to distinguish real images from those produced by a particular model. By analyzing the classifier’s decisions with saliency maps, the authors show that the maps reveal model‑specific artifacts rather than a universal notion of fakeness, and that combining classification accuracy with saliency information yields a composite score that aligns with human judgments. This approach avoids reliance on large pretrained backbones and works solely with real and generated samples from the evaluated model.
By Eleonora Ristori, Luca Bindini, Marco Lippi, Paolo Frasconi
TAN (Tissue‑Aware Network) is a lightweight convolutional encoder that processes tissue‑conforming superpixels in their native shape, avoiding background leakage from padding or masking. It achieves accuracy on the BACH histology dataset comparable to a 303‑million‑parameter foundation model while using only 0.41 M parameters and 1,190× fewer FLOPs, and its embeddings are invariant to the choice of fill value. On the BRACS dataset, TAN performs close to a ResNet18 trained under the same protocol and outperforms standard masked‑crop processing with less than half the computational cost.
By Jan Herschel, Anselm von Gladiss
Video Diffusion Avatars (VDAs) fine‑tune video diffusion models to generate photo‑realistic, multi‑view consistent avatars from a single image, allowing precise control over camera viewpoint, facial expression, and upper‑body pose. VDAs replace traditional 3D morphable models with a learned expression space, overcoming viewpoint bias through disentanglement and enabling independent control of expression, pose, and viewpoint via proxy‑mesh renderings. The method trains efficiently—34 hours on a single A100—and outperforms state‑of‑the‑art single‑image 3D avatars, improving novel‑view reenactment metrics and producing physically plausible hair dynamics, shadows, reflections, and refractions.
By Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen, Matthias Nie{\ss}ner
The paper introduces StopWatch, a method that explicitly grounds the goal in object-goal navigation tasks. By addressing the dominant failure of goal grounding in fine‑tuned vision‑language models, StopWatch achieves state‑of‑the‑art success rates on six benchmarks across three scene collections and corrects annotation defects in these datasets. The authors demonstrate that with accurate goal grounding, their policy nearly matches the expert demonstrations, highlighting perception as the remaining bottleneck.
By Xinhao Liu, Yunsheng Tian, Yuheng Qiu, Pieter Abbeel, Xi Chen, Guanya Shi, Chen Feng
The paper introduces LGGS, a deep‑learning framework that uses prior longitudinal MRI scans to guide accelerated image reconstruction through a learned, phase‑invariant circulant attention mechanism. This approach incorporates group sparsity principles to handle misregistration and phase variation, achieving state‑of‑the‑art performance on longitudinal T1‑weighted brain MRI data with fewer parameters than existing methods. The authors also show that even accelerated prior images can serve as effective guides with minimal quality loss.
By Nikola Janju\v{s}evi\'{c}, Hersh Chandarana, Yao Wang, Li Feng
MambaAct introduces a compact Mamba-based action expert to replace the transformer-based expert in flow‑matching vision‑language‑action models, preserving the pretrained visual‑language prefix encoder while reducing the action expert size by tenfold. The new design employs a lightweight Context Compressor to embed vision‑language conditioning into each Mamba block, eliminating repeated token‑level cross‑attention. Experiments show a 64.8 % reduction in denoising latency on LIBERO and 58.8 % in real‑world deployment, with a modest 1.08 % increase in mean success rate across multiple fine‑tuning seeds.
By Khuyen Pham, Tran Nguyen Le, Edoardo Fazzari, Cesare Stefanini
GatewayBench is a new evaluation framework that systematically studies how the design of Model Context Protocol (MCP) gateways influences AI agent capability, efficiency, and security. It categorizes gateways into flattened and hierarchical architectures and tests them on 524 tasks across six domains using a pool of 11,118 real-world tools from 251 MCP servers, including task-irrelevant tools to simulate large-scale deployments. Experiments on seven widely deployed gateways reveal significant differences in task completion, tool-use efficiency, and attack resilience, showing strong interactions with the underlying model and agent harness.
By Yitian Zheng, Zhaorun Chen, Haoyu Li, Bo Li
TransNSNet is a Transformer‑enhanced multi‑scale adversarial framework designed for CT‑to‑PET image synthesis. The study curated a paired CT‑PET dataset from a public NSCLC Radiogenomics collection and evaluated the framework against several baselines, reporting that TransNSNet achieved the best mean performance in nine of ten metrics while maintaining a smaller parameter footprint. Statistical analysis found no significant superiority over other models after correction, and the authors caution that synthesized PET‑like images should not replace clinically acquired PET scans.
By Shaojun Li, Xiaoyu Deng, Ikuo Keshi
The paper introduces UniDecouple, a selective decoupling framework for continual learning in unified multimodal models that addresses Task Objective Conflict (TOC). By employing a Task-Aware Mixture-of-Experts (TA‑MoE) architecture, it retains shared attention layers while adding task‑specific feed‑forward network experts to reduce optimization conflicts. Hierarchical expert routing and hybrid collaboration further balance task specialization with cross‑task knowledge sharing, and a two‑stage training strategy enhances both multimodal understanding and image generation performance.
By Jiaxing Zhang, Hao Tang
The paper introduces GazeCLIP, a gaze‑guided CLIP framework that uses adaptive‑enhanced fine‑grained language prompts for deepfake attribution and detection (DFAD). It presents a new benchmark evaluating DFAD on novel generative methods such as diffusion and flow models, and proposes a gaze‑aware image encoder (GIE) that fuses gaze prompts with image embeddings to capture global forgery patterns. A language refinement encoder (LRE) further refines language embeddings for precise vision‑language matching, achieving state‑of‑the‑art performance with 6.56% higher accuracy and 5.32% higher AUC on the benchmark.
By Yaning Zhang, Linlin Shen, Zitong Yu, Chunjie Ma, Zan Gao
LDDR (Linear DPP Dynamic Resolution) is a training‑free, budget‑aware video frame sampling framework designed for multimodal large language models (MLLMs). It uses query‑aware Determinantal Point Process (DPP) selection in a task‑conditioned feature space, achieving a 3× runtime speedup over standard DPP baselines. By assigning a non‑uniform token budget—allocating more tokens to informative, non‑redundant frames and downscaling or pruning less useful ones—LDDR improves performance by 2.5 points in low‑budget settings and 1.6 points in high‑budget scenarios across four video benchmarks and multiple MLLM families.
By Jingfeng Chen, Jiawen Qian, Wendi Deng, Yinuo Guo, Jiaqi Yu, Sicong Leng, Raghuveer Thirukovalluru, Bhuwan Dhingra
TSA: Temporal Slot Activation for Persistent Object-Centric Video Representation proposes a lightweight mechanism that learns per-slot, per-frame activation scores to control slot lifecycle in unsupervised video object-centric learning. By gating slot updates and decoder participation when objects are absent or occluded, TSA reduces state drift and reconstruction interference, improving object decomposition and temporal identity preservation. The method is evaluated on MOVi-C, MOVi-E, YTVIS-HQ, and OVIS benchmarks, showing significant gains, especially on long, heavily occluded videos.
By Duc Nguyen, Sieu Tran, Hao Vo, Khoa Vo, Duy Minh Ho Nguyen, Nghi D. Q. Bui, Anh Nguyen, Long Mai, Ngan Le
BG-REAL is a public benchmark for detecting and localizing background manipulations in images, built from Open Images V7 with 7,000 processed samples across 1,200 source groups. It includes 6,000 real-data anchored samples, 1,000 synthetic controls, six edit families, matched authentic controls, and extensive quality‑control measures. The release offers a construction pipeline, evaluation protocol, and three external baseline results, highlighting false‑positive rates due to re‑encoding artifacts.
By Bugra Alperen Uluirmak, Rifat Kurban
The paper introduces the phase‑induced amplitude apodisation Zernike wavefront sensor (PIAA‑ZWFS), a modified ZWFS that uses lossless pupil apodisation to concentrate starlight in the focal plane. Using a differentiable modelling framework and Bayesian experimental design, the authors optimise the sensor to minimise the variance of a maximum‑likelihood estimator in the high‑Strehl regime. Simulations show that the PIAA‑ZWFS achieves state‑of‑the‑art performance across various apertures, bandwidths, photon fluxes, and source sizes, closing the gap to the fundamental limit by a factor of 10 (or 2.5) compared to conventional or optimised ZWFS, and outperforming ideal point‑source sensors for stellar diameters larger than 0.8 λ/D without sacrificing dynamic range.
By A. K. Taras, S. Y. Haffert, L. Desdoigts