著者: Marc A. Kastner, Shin'ichi Satoh

あらすじ:

An understanding of human perception and sentiment can help with multimedia tasks like video recommendations. Having a model of attraction of videos would be promising to find other related videos giving the viewer a similar sentiment. On social media platforms, the sentiment of videos can be a crucial factor for the popularity of a video and its similarity to others. In this research, we propose a framework to model the sentiment of social media videos by first analyzing the sentiment of its respective user comments. We decide a sentiment annotation for each video in our base video dataset through text sentiment analysis. From this, we train a model towards the prediction of viewer sentiment by analyzing audio-visual features. A preliminary study can show promising performance in predicting the comment sentiment annotations from audio-visual features.

種類: Talk at Meeting of the Technical Committee on Media Experience and Virtual Environment, MVE (メディアエクスペリエンス・バーチャル環境基礎研究会)

日付: September 2020