Bollywood Movie Corpus for Text, Images and Videos

Published in arXiv, 2017

A large multimodal dataset (text, images, video) designed for computational media analysis, enabling research on representation, fairness, and cultural analytics in film.

Recommended citation: N. Madaan, S. Mehta, M. Saxena, A. Aggarwal, T. Agrawaal, V. Malhotra. Bollywood Movie Corpus for Text, Images and Videos. arXiv preprint arXiv:1710.04142, 2017.
Download Paper