cs.CV · 2026-06-15 · No. 24

Computer Vision and Pattern Recognition, 2026-06-15.

17 new papers in cs.CV. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

17 entries
  1. 01

    Gaze Heads: How VLMs Look at What They Describe

    Rohit Gandikota, David Bau

    cs.CV · cs.CL · cs.LG

    How a vision-language model internally solves the task of describing an image is far from obvious. We find that the model develops a specific mechanism for this: a small set of attention heads in its language-model backbone, which we call gaze heads, whose attention tracks the image region the model is currently describing. We find them with a simple correlation score from a few forward passes, using comic strips as a controlled testbed where...

    arxiv.org/abs/2606.14703 · PDF

  2. 02

    ClinHallu: A Benchmark for Diagnosing Stage-Wise Hallucinations in Medical MLLM Reasoning

    Sicheng Yang, Hangjie Yuan, Wenjun Zhang, Jinwang Wang, Yichen Qian, Weihua Chen, Fan Wang, Lei Zhu

    cs.CV · cs.AI · cs.CL

    Building trustworthy medical multimodal large language models (MLLMs) is critical for reliable clinical decision support. Existing medical hallucination benchmarks mainly focus on data collection, but often ignore where hallucinations originate within the reasoning process. We find that hallucination sources vary across samples: errors may arise from visual misrecognition, incorrect medical knowledge recall, or flawed reasoning integration....

    arxiv.org/abs/2606.14697 · PDF

  3. 03

    CottonLeafVision: An Explainable and Robust Deep Learning Framework for Cotton Leaf Disease Classification

    Rafi Ahamed, Md. Abir Rahman, Tasnia Tarannum Roza, Munaia Jannat Easha, Md. Asif Khan, Sudeepta Mandal

    cs.CV · cs.AI

    Globally, cotton is a highly economically beneficial crop, as the textile industry heavily depends on it. So, the precise identification and detection of cotton leaf disease is crucial for economic stability. The development goal of "CottonLeafVision" is to accurately classify and detect cotton leaf disease. With this goal, we have evaluated multiple pretrained Deep Convolutional Neural Networks, including DenseNet201, InceptionV3, and VGG19...

    arxiv.org/abs/2606.14686 · PDF

  4. 04

    HumP-KD: A Hybrid Uncertainty-Aware Multi-Stage Progressive Knowledge Distillation Framework for Efficient Fire Classification

    Mohammed Arif Mainuddin, Najifa Tabassum, Omar Ibne Shahid, Riasat Khan

    cs.CV · cs.LG

    Real-time fire classification systems require models that are simultaneously accurate, computationally efficient, and deployable on resource-constrained hardware. This work proposes \textbf{HumP-KD}, a Hybrid Uncertainty-aware Multi-stage Progressive Knowledge Distillation framework for efficient fire classification. Two datasets, FlameVision and Dataset-II, containing 8,600 and 31,309 images, are used. Various CNN and transformer baselines...

    arxiv.org/abs/2606.14684 · PDF

  5. 05

    Giving AI a Headache: Acoustic Adversarial Attacks to Computer Vision Applications

    Nicole Villavicencio-Garduño, Maksim Ekin Eren, Milo Prisbrey, Ben Migliori, Michael Teti

    cs.CV · cs.AI

    Artificial Intelligence (AI) is increasingly used to automate a variety of real-world computer vision (CV) applications, such as autonomous vehicle control, facial recognition, and security cameras. Recent research has shown that acoustic vibration can induce real physical motion in cameras, interfering with their internal stabilization mechanisms. Because the motion falls outside the conditions the stabilization system was designed to...

    arxiv.org/abs/2606.14658 · PDF

  6. 06

    NEST3D: A High-Resolution Multimodal Dataset of Sociable Weaver Tree Nests

    Constanza A. Molina Catricheo, Simon Boeder, Ting-Jia Guo, Giacomo May, Clément Berthelot, Devis Tuia, Friedrich...

    cs.CV · cs.LG

    Sociable weaver nests function as complex ecological structures offering thermoregulatory microhabitats and sustaining diverse species; however, datasets used in prior studies lack fine-grained 3D structural detail. Producing usable and accurate 3D weaver nest data is challenging due to their irregular geometry and integration with complex host vegetation. We bridge this gap with an open-access, 1.4 TB multimodal drone dataset of 104...

    arxiv.org/abs/2606.14562 · PDF

  7. 07

    Rethinking Global Average Pooling: Your Classifier Is Secretly a Multi-Instance Learner

    Aray Karjauv

    cs.CV · cs.AI

    Modern image classifiers widely adopt global average pooling (GAP) followed by a linear classification head. This linearity ensures that the image-level logits equal the average of logits obtained by applying the classification head pointwise to the feature grid prior to GAP. Consequently, standard classifiers may inherently retain spatial class evidence that remains recoverable even when the image-level prediction is incorrect. This...

    arxiv.org/abs/2606.14555 · PDF

  8. 08

    What Drives Test-Time Adaptation for CLIP? A Controlled Empirical Study from an Update Perspective

    Jiazhen Huang, Xiao Chen, Zhiming Liu, Yaru Sun, Jingyan Jiang, Zhi Wang

    cs.CV · cs.LG

    Vision-Language Models (VLMs) such as CLIP have become a standard backbone for open-vocabulary recognition, yet their zero-shot predictions remain vulnerable to distribution shifts encountered at deployment. Test-Time Adaptation (TTA) has recently been extended to CLIP as a lightweight solution, leading to a rapidly growing body of TTA4CLIP methods. However, empirical progress in this area has largely outpaced our understanding of what truly...

    arxiv.org/abs/2606.14299 · PDF

  9. 09

    Pix2Pix-Hybrid: Structure-Guided Conditional Synthesis of Hajj Crowd Images with Multi-Channel Conditioning and Weak Attribute Supervision

    Amirah F. Alshammari, Bander A. Alzahrani, Nahed A. Alowidi

    cs.CV · cs.AI

    Developing accurate crowd-counting models for Hajj pilgrimage scenes remains challenging because domain-specific annotated images are scarce and data collection during large gatherings raises privacy concerns. To address these limitations, this paper proposes Pix2Pix-Hybrid (P2P-H), a hybrid conditional GAN for structure-guided Hajj crowd-image synthesis and data augmentation. P2P-H builds on Pix2Pix and employs a U-Net generator conditioned...

    arxiv.org/abs/2606.14297 · PDF

  10. 10

    Hybrid Classical-Quantum (HCQ) Alzheimer's Classification via Supervised $β$-VAE and Quantum Kernels

    Tia Tiwari, Vamshi Krishna Kancharla, Neelam Sinha

    cs.CV · cs.LG

    This paper presents a two-stage Hybrid Classical-Quantum (HCQ) pipeline for binary Alzheimer's disease (AD) classification from 3D T1-weighted structural MRI volumes, where the classical and quantum components are designed to complement each other rather than operate independently. A supervised 3D $β$-variational autoencoder (VAE) is trained end-to-end under voxel-wise reconstruction, KL-divergence, and focal classification losses that...

    arxiv.org/abs/2606.14194 · PDF

  11. 11

    Conditioning Matters: Stabilizing Inversion and Attention in Diffusion Image Editing

    Zheyuan Zhan, Hongchen Li, Can Wang, Yinfei Ma, Mingzhen Huang, Ruoshi Bai, Jiawei Chen, Siwei Lyu, Defang Chen

    cs.CV · cs.AI

    Inversion-based image editing offers flexible and training-free control but still struggles with inversion accuracy and the trade-off between editing fidelity and background preservation. While recent methods improve inversion formulations or attention interactions, the role of textual conditioning in shaping diffusion dynamics and editing behavior remains underexplored. We show both empirically and theoretically that the precision of textual...

    arxiv.org/abs/2606.14125 · PDF

  12. 12

    FEMOT: Multi-Object Tracking using Frame and Event Cameras

    Shiao Wang, Xiao Wang, Chao Wang, Yitao Li, Menghao Liu, Bo Jiang, Yaowei Wang, Yonghong Tian, Jin Tang

    cs.CV · cs.AI

    Conventional RGB cameras have been widely used in multi-object tracking due to their ability to capture rich appearance and semantic information. However, their performance is often degraded under complex real-world challenges, such as motion blur, low illumination, and overexposure. Bio-inspired event cameras offer high temporal resolution and high dynamic range, providing complementary cues under extreme scenarios. Nevertheless, RGB-event...

    arxiv.org/abs/2606.14094 · PDF

  13. 13

    Clay-CNN Hybrids: Leveraging Geo-Foundational Models as Auxiliary Context for Landslide Detection

    Huong Binh Vu

    cs.CV · cs.AI · cs.LG · eess.IV

    Rapid post-event landslide mapping is essential for disaster response but remains difficult to automate due to extreme class imbalance. This study evaluates whether Clay v1.5, a Geo-Foundational Model (GFM), can improve pixel-level landslide segmentation on the Landslide4Sense (L4S) benchmark, which contains 3,799 training chips with 14 Sentinel-2 and terrain bands and approximately 2% positive pixels. We compare three strategies: Clay as the...

    arxiv.org/abs/2606.14081 · PDF

  14. 14

    RT-VLA: Real-Time Vision-Language-Action Models via Knowledge Distillation

    Xiangyu Huang, Zhenlin Hua, Han Zhou, Shounak Sural, Ragunathan Rajkumar

    cs.CV · cs.LG · cs.RO

    Vision-Language-Action (VLA) models have shown strong potential for end-to-end autonomous driving by jointly modeling visual perception, language reasoning, explainability and action prediction. However, their large vision-language backbones and reasoning modules introduce substantial inference latency and thereby prevent their deployment in the unforgiving reality of the road networks. We propose RT-VLA, a lightweight, distilled VLA model...

    arxiv.org/abs/2606.14010 · PDF

  15. 15

    Self-Evolving Visual Questioner

    Yijun Liang, Hengguang Zhou, Ming Li, Lichen Li, Cho-Jui Hsieh, Tianyi Zhou

    cs.CV · cs.LG

    Vision-language models (VLMs) are typically trained as passive answerers, while their ability to actively ask diverse, non-trivial, visual-centric and grounded questions remains underexplored. Existing visual questioners' performance is bottlenecked by the availability of high-quality training data or the cost of curating them. We show that a VLM can continuously improve itself as a visual questioner without any external supervision. We...

    arxiv.org/abs/2606.13929 · PDF

  16. 16

    HiLo-Token: Input-Adaptive High-Low Frequency Token Compression for Efficient Image Editing

    Haoran You, Yotam Nitzan, Lingzhi Zhang, Yifan Gong, Mang-Tik Chiu, Connelly Barnes, Yan Kang, Yuqian Zhou, Eli...

    cs.CV · cs.AI

    Creative image editing tools, such as Photoshop's Remove or Generative Fill buttons, are central to everyday customer use and account for a major share of traffic in Photoshop and Lightroom. However, current generative AI models face significant latency challenges, which become even more pronounced when transitioning from convolution-based U-Nets to Diffusion Transformers (DiTs). In our evaluation on hundreds of representative image editing...

    arxiv.org/abs/2606.13898 · PDF

  17. 17

    How do Self-Supervised Remote Sensing Vision Models Transfer to Downstream Tasks?

    Julia Romero, Qin Lv, Morteza Karimzadeh

    cs.CV · cs.AI

    Self-supervised geospatial foundation models (GeoFMs) learn transferable representations from remote sensing data, but their downstream behavior is difficult to characterize. We study six representative GeoFMs spanning joint-embedding, reconstruction, and multimodal pretraining families, and evaluate transfer across classification, regression, and segmentation benchmarks under different label availability and downstream pipelines. We find...

    arxiv.org/abs/2606.13896 · PDF

This edition is part of The Daily Abstract — cs.CV archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.