Can SAM Count Anything? An Empirical Study on SAM Counting

Zhiheng Ma, Xiaopeng Hong, Qinnan Shangguan
4/21/2023
cs.CVcs.AI

Official code: Vision-Intelligence-and-Robots-Group/count-anything (95 stars, MIT, last commit 2023-04-25).

Abstract

Meta AI recently released the Segment Anything model (SAM), which has garnered attention due to its impressive performance in class-agnostic segmenting. In this study, we explore the use of SAM for the challenging task of few-shot object counting, which involves counting objects of an unseen category by providing a few bounding boxes of examples. We compare SAM's performance with other few-shot counting methods and find that it is currently unsatisfactory without further fine-tuning, particularly for small and crowded objects. Code can be found at \url{https://github.com/Vision-Intelligence-and-Robots-Group/count-anything}.

View on arXivView PDF

Code Implementations(1)

Vision-Intelligence-and-Robots-Group/count-anythingOfficial100%
MIT

an empirical study on few-shot counting using segment anything (SAM)

9511Python, Jupyter NotebookApr 18, 20233 years agoMIT

Similar Papers

Few-shot Object Counting and Detection

88% match1 repo852022

Point, Segment and Count: A Generalized Framework for Object Counting

87% match1 repo1252023

T-Rex: Counting by Visual Prompting

86% match2023

Referring Expression Counting

85% match2024

Object Counting with GPT-4o and GPT-5: A Comparative Study

85% match2025

Can Current AI Models Count What We Mean, Not What They See? A Benchmark and Systematic Evaluation

85% match2025

Cite this paper

@article{ma2023count,
  title  = {Can SAM Count Anything? An Empirical Study on SAM Counting},
  author = {Zhiheng Ma and Xiaopeng Hong and Qinnan Shangguan},
  year   = {2023},
  eprint = {2304.10817},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2304.10817v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2304.10817.svg)](https://codeofpaper.com/paper/2304.10817)

Discussion