FLEXPART framework enables text-controlled 3D mesh segmentation

Researchers have introduced FLEXPART, a novel framework designed for text-controlled 3D mesh segmentation. This system allows users to specify desired parts of a 3D asset using natural language prompts, which can include semantic part names along with location and shape descriptions. To facilitate this, FLEXPART utilizes a Vision-Language Model (VLM) pipeline to create a large dataset of multi-granularity partitions and corresponding prompts. The framework generates a view-independent part-identity field in 3D space, directly assigning part IDs to mesh faces. FLEXPART has demonstrated superior performance, outperforming strong baselines by significant margins in mean Intersection over Union (mIoU) metrics. AI

IMPACT This framework could streamline 3D asset manipulation and analysis by enabling intuitive, text-based control over part segmentation.

RANK_REASON The cluster describes a new research paper introducing a novel framework for 3D mesh segmentation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• FLEXPART
• FLEXPARTBENCH
• Hugging Face
• vision-language model

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →