LLMSeg: Training-Free Text-Prompted Image Segmentation with B-Splines
We present LLMSeg, a training-free pipeline that turns a vision-capable LLM into a text-prompted image segmenter.
ProofPaper ↗
Key points
- Promptable segmentation models such as SAM 3 achieve remarkable open-vocabulary performance but require dedicated architectures, large-scale annotation engines, and task-specific training.
- We ask a different question: can a general-purpose multimodal large language model (MLLM) segment images directly, without an added segmentation head, mask decoder, or training?
- LLMSeg decomposes segmentation into three LLM-driven stages.
- We evaluate LLMSeg across ten referring and reasoning segmentation splits (ReasonSeg and RefCOCO/+/g).
Sources (1)
- [1]LLMSeg: Training-Free Text-Prompted Image Segmentation with B-SplinesarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 08:35 AM
We present LLMSeg, a training-free pipeline that turns a vision-capable LLM into a text-prompted image segmenter.
Promptable segmentation models such as SAM 3 achieve remarkable open-vocabulary performance but require dedicated architectures, large-scale annotation engines, and task-specific training.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 21, 2026Steering Black-Box Large Language Model Agents with Pitfall Anticipation World Model
- Oct 20, 2026Attention-based Classification for Indirect Prompt Injection
- Oct 20, 2026I built an RK3588 NPU runtime without any proprietary inference libraries: Qwen3.5-4B at ~10 tok/s, ~2x faster than vendor SDK.
- Oct 20, 2026RoboJudge: Multimodal Language Models as Judges for Embodied Video Generation
- Oct 19, 2026We built our own inference engine so a 32B model could keep 262K context inside a 16 GB GPU budget. Here's an uncut run.
- Oct 19, 2026I Can See Everything