MOUD Dataset


The Multimodal Opinion Utterances Dataset (MOUD) Dataset consists of 80 product review videos in Spanish. Videos were found on the YouTube search page using the following keywords (translated to English): my favorite products, non recommended products, my favorite perfumes, recommended movies), non recommended movies, recommended books and non recommended books. Each video consists of multiple segments labeled to display positive, negative or neutral sentiment.

MOUD dataset is available through CMU Multimodal Data SDK on GitHub: