Researcher, Toyota Motor Europe
1 paper at NeurIPS 2025
We propose LangHOPS, the first Multimodal Large Language Model (MLLM)-based framework for open-vocabulary object–part instance segmentation.