Scholarlink

In this paper, a method to efficiently generate multi-task, dense subitizing detection diagrams is proposed. It is inspired by the billiards game and implemented with use of intentionally simplified two-dimensional physics simulation. The observed advantage of proposed Billiards Subitizing Diagrams Generation (BSDG) over greedy search and two variants of simulated annealing suggests that it can serve as a search method useful in solving many problems which can be related with billiards and other similar games. Proposed solution can be further improved by the use of computation parallelization and possibly by relating it to theoretical foundations of the circle packaging problem. But even in its current form, BSDG can be used to generate valuable, large data sets of images dedicated to the analysis of the subitizing effect in machine learning models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Billiards-Based Generation of Multi-task, Dense Subitizing Detection Diagrams

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Maciej Huk

摘要

In this paper, a method to efficiently generate multi-task, dense subitizing detection diagrams is proposed. It is inspired by the billiards game and implemented with use of intentionally simplified two-dimensional physics simulation. The observed advantage of proposed Billiards Subitizing Diagrams Generation (BSDG) over greedy search and two variants of simulated annealing suggests that it can serve as a search method useful in solving many problems which can be related with billiards and other similar games. Proposed solution can be further improved by the use of computation parallelization and possibly by relating it to theoretical foundations of the circle packaging problem. But even in its current form, BSDG can be used to generate valuable, large data sets of images dedicated to the analysis of the subitizing effect in machine learning models.