PulseAugur
EN
LIVE 20:10:30

New model HieraCount tackles multi-grained object counting with large dataset

Researchers have introduced a new approach to open-world object counting, addressing the brittleness of current methods by redefining the problem as multi-grained counting. This method explicitly defines the desired granularity of counting through visual exemplars and fine-grained text, including negative prompts, across five distinct levels. To support this, they developed an automatic data-scaling pipeline using 3D synthesis and VLM-based filtering to create KubriCount, the largest dataset for counting tasks, and trained a model called HieraCount that significantly improves accuracy and generalization. AI

IMPACT Improves object counting accuracy and generalization by explicitly handling counting granularity, potentially benefiting applications in computer vision and robotics.

RANK_REASON The cluster describes a new research paper introducing a novel model and dataset for object counting. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Hugging Face Daily Papers →

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

New model HieraCount tackles multi-grained object counting with large dataset

COVERAGE [1]

  1. Hugging Face Daily Papers TIER_1 (TL) ·

    Count Anything at Any Granularity

    Open-world object counting remains brittle: despite rapid advances in vision-language models (VLMs), reliably counting the objects a user intends is far from solved. We argue that a central reason is that counting granularity is left implicit; users may refer to a specific identi…