Document Type

Article

Domestic/International

Domestic

Location

New London, Connecticut, United States

Publication Date

2026

Abstract

This reflection describes an 8 week summer research program conducted through the Connecticut College Summer Science Research Institute under the supervision of Professor Ozgur Izmirli. The project involved building a pipeline to automatically detect bow direction in violin performance videos and annotate the corresponding sheet music, using optical flow to track bowing motion and a machine learning classifier to label each video frame. Ground truth annotations were generated across violin performance pieces to train and evaluate the system.

Download
COinS
 

The views expressed in this paper are solely those of the author.