arXiv Computer Vision September 22, 2026 By Chaoqian Mu, Wenhao Wu, Zichen Liang, Jiaxu Li, Lijun Wang, Yifan Wang, Huchuan Lu

MinCU: A Fine-Grained Benchmark for Grounded Minimal-Change Understanding in Image Pairs

Read the original on arXiv Computer Vision →

MinCU is a new benchmark for grounded minimal‑change understanding that presents pairs of near‑identical images differing by a single atomic variation in object category, attribute, count, or spatial position. Models are evaluated on their ability to describe the change, localize the changed region, and identify the changed entity. The authors also introduce SG‑ISA, a structured autoregressive method that decomposes the task into a Think‑Locate‑Describe sequence, showing that fine‑tuning with SG‑ISA improves both grounding accuracy and description quality while reducing reasoning‑token overhead.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

arXiv AI

Sep 10

VDiff-Bench: A Challenging Benchmark for Fine-Grained Image Difference Identification

arXiv:2609.06245v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) perform strongly on general visual understanding tasks such as visual question answering, yet they often str...

By Yixin Wan, Tianle Zheng, Kai-Wei Chang

llms computer-vision nlp multimodal benchmarks

arXiv AI

2d ago

Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes

arXiv:2610.02117v1 Announce Type: cross Abstract: On-policy self-distillation has recently emerged as an effective approach for improving language-model reasoning by supervising students with a froze...

By Sophia Sirko-Galouchenko, Monika Wysoczanska, Andrei Bursuc, Nicolas Thome, Spyros Gidaris

llms efficiency multimodal benchmarks

arXiv AI

Aug 12

Grounded Post-Training with Hard Examples for Reducing Hallucination in Multimodal Large Language Models

arXiv:2605. 16411v3 Announce Type: replace-cross Abstract: Hallucination remains a fundamental challenge in vision-language models (VLMs), where autoregressive generation may produce linguistically plausible yet physically inconsistent or visually ungrounded responses due to likelihood maximization under joint probabilistic modeling.

By Qinwu Xu

llms computer-vision reinforcement-learning multimodal benchmarks safety

arXiv AI

Jun 30

CCRC: A Change-Aware Captioning and Reasoning Chain for Image Change Captioning and Segmentation

arXiv:2606. 28724v1 Announce Type: cross Abstract: Understanding and localizing subtle changes between paired images is critical for tasks such as surveillance and image editing.

By Jinhong Hu, Xiaoping Wang, Shuyin Huang, Guojin Zhong, Kaitai Liu, Kai Lu

llms computer-vision multimodal benchmarks

arXiv AI

Jun 16

Thinking with Visual Grounding

arXiv:2606. 16122v1 Announce Type: new Abstract: Visual thinking should not only sound right; it should show its evidence.

By Junkai Zhang, Yihe Deng, Kai-Wei Chang, Wei Wang

llms agents reinforcement-learning multimodal benchmarks

arXiv Machine Learning

Aug 12

MultiModal Code-Switching: Interleaving Visual Objects into Language for Explicit Object-Level Alignment

arXiv:2608. 11167v1 Announce Type: cross Abstract: Existing Multimodal Large Language Models (MLLMs) predominantly rely on image-text pairs for modality alignment pretraining, mapping global image representations to long textual descriptions.

By Changhao Xiang, Shangyu Xing, Zhen Wu, Jianbing Zhang, Xinyu Dai

llms multimodal safety

One email a morning, machine-written

Related stories

VDiff-Bench: A Challenging Benchmark for Fine-Grained Image Difference Identification

Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes

Grounded Post-Training with Hard Examples for Reducing Hallucination in Multimodal Large Language Models

CCRC: A Change-Aware Captioning and Reasoning Chain for Image Change Captioning and Segmentation

Thinking with Visual Grounding

MultiModal Code-Switching: Interleaving Visual Objects into Language for Explicit Object-Level Alignment