Fingertips tracking algorithm for guitarist based on temporal grouping and pattern analysis

Zhao Wang*, Jun Ohya

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)


This paper proposes a temporal Grouping and pattern analysis-based algorithm that could track the fingertips of guitarists during their guitar playing towards the actualization of the automatic guitar fingering recognition system. First a machine learning-based Bayesian Pixel Classifier is used to segment the hand area on the test data. Then, the probability map of fingertip is generated on the segmentation results by counting the voting numbers of the Template Matching and Reversed Hough Transform. Furthermore, a temporal Grouping algorithm, which is a geometry analysis for consecutive frames, is applied to removal noise and group the same fingertips (index finger, middle finger, ring finger, little finger). Then, a data association algorithm is utilized to associate 4 tracked fingers (index finger, middle finger, ring finger, little finger) with their correspondent tracked results frame by frame. Finally, particles are distributed only between the associated fingertip candidates to track the fingertips of guitarist effectively. The experimental result demonstrates that this fingertip tracking algorithm is robust enough for tracking fingertips (1) without any constrains such us color marker; (2) under the complex contexts, such us complicated background, different illumination conditions, (3) with the high tracking accuracy (mean error 3.36 pixels for four fingertips).

Original languageEnglish
Title of host publicationComputer Vision - ACCV 2016 Workshops, ACCV 2016 International Workshops, Revised Selected Papers
EditorsChu-Song Chen, Kai-Kuang Ma, Jiwen Lu
PublisherSpringer Verlag
Number of pages15
ISBN (Print)9783319545257
Publication statusPublished - 2017
Event13th Asian Conference on Computer Vision, ACCV 2016 - Taipei, Taiwan, Province of China
Duration: 2016 Nov 202016 Nov 24

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10118 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other13th Asian Conference on Computer Vision, ACCV 2016
Country/TerritoryTaiwan, Province of China
City Taipei

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Fingertips tracking algorithm for guitarist based on temporal grouping and pattern analysis'. Together they form a unique fingerprint.

Cite this