An Open Source Quantitative Evaluation Framework for Automatic Video Summarization Algorithms

ARIEL I. DIAZ; ANÍBAL M. BLANCO; ADRIA ́N ROSTAGNO; JAVIER IPARRAGUIRRE; LEANDRO BALMACEDA; SANTIAGO L. AGGIO

Salta

Simposio; SAIV - Simposio Argentino de Imágenes y Visión; 2019

JAIIO

The creation, consumption, and manipulation of video play a centralrole in everyday life as the amount of video data is growing at an exponentialrate. Video summarization consists on producing a condensed output from avideo that allows humans to rapidly understand and browse the content of theoriginal source. Although there are several evaluation approaches proposed inthe literature, multiple challenges make the quantitative evaluation of asummarization a complex process. In this paper we present a completely openvideo summarization evaluation framework that is compatible with existingdatasets and published results. Standard metrics are considered and a new metricthat captures unbalanced-class video summarization evaluation is proposed. Twolegacy datasets are integrated in a standard format. Finally, new quantitativeresults based on already published algorithms are presented.