Teaching AI systems to understand intent and affordances through conversational image segmentation at scale.
Conversational image segmentation grounds abstract, intent-driven concepts into pixel-accurate masks. Prior work on referring image grounding focuses on categorical and spatial queries (e.g., "left-most apple") and overlooks functional and physical reasoning (e.g., "where can I safely store the knife?"). We address this gap and introduce Conversational Image Segmentation (CIS) and ConverSeg, a benchmark spanning entities, spatial relations, intent, affordances, functions, safety, and physical reasoning. We also present ConverSeg-Net, which fuses strong segmentation priors with language understanding, and an AI-powered data engine that generates prompt-mask pairs without human supervision. We show that current language-guided segmentation models are inadequate for CIS, while ConverSeg-Net trained on our data engine achieves significant gains on ConverSeg and maintains strong performance on existing language-guided segmentation benchmarks. Project webpage: https://glab-caltech.github.io/converseg/
Conversational image segmentation enables AI systems to understand functional properties and user intent beyond basic object recognition.
A five-stage automated data engine generates pixel-accurate masks with natural language prompts without human supervision at scale.
Five concept families—entities, spatial relationships, interactions, affordances, and physics—capture the full spectrum of conversational reasoning about images.
Traditional vision-language models struggle with complex reasoning about affordances, physics constraints, and functional properties in real-world scenarios.
Creating pixel-accurate masks paired with reasoning-heavy natural language prompts is prohibitively expensive with human annotation alone.
Jingxuan Wei, Siyuan Li +22
The construction of World Models capable of learning, simulating, and reasoning about objective physical laws constitutes a foundational challenge in the pursuit of Artificial General Intelligence. Re...
Haifeng Wang, Hua Wu +436
In this report, we introduce ERNIE 5.0, a natively autoregressive foundation model desinged for unified multimodal understanding and generation across text, image, video, and audio. All modalities are...
Ailin Huang, Chengyuan Yao +91
We present STEP3-VL-10B, a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. STEP3-VL-10B is realized t...
Yu Wang, Yi Wang +5
As hubs of human activity, urban surfaces consist of a wealth of semantic entities. Segmenting these various entities from satellite imagery is crucial for a range of downstream applications. Current ...
Hongrui Jia, Chaoya Jiang +2
As Large Multimodal Models (LMMs) scale up and reinforcement learning (RL) methods mature, LMMs have made notable progress in complex reasoning and decision making. Yet training still relies on static...
