We discuss a novel deep affect-based movie trailer classification framework. Affective video content analysis has emerged as one of the most challenging and essential research tasks as it aims to analyze the emotions elicited by videos automatically. However, little progress has been achieved in this field due to the enigmatic nature of emotions. This widens the gap between the human affective state and the structure of the video. In this paper, we propose a novel deep affect based movie trailer classification framework. We also develop an EmoGDB dataset, which contains 100 Bollywood movie trailers annotated with popular movie genres: Action, Comedy, Drama, Horror, Romance, Thriller, and six different types of induced emotions: Anger, Fear, Happy, Neutral, Sad, Surprise. The affect-based features are learned via ILDNet architecture trained on the EmoGDB dataset.
Our work aims to analyze the relationship between the emotions elicited by the movie trailers and how they contribute in solving the multi-label genre classification problem. The proposed novel framework is validated by performing cross-dataset testing on three large scale datasets, namely LMTD-9, MMTF- 14K, and ML-25M datasets. Extensive experiments show that the proposed algorithm outperforms all the state-of-the-art methods significantly, as reported by the precision, recall, F1 score, precision–recall curves (PRC), and area under the PRC evaluation metrics.
• Development of a unified framework of deep networks for movie genre classification.
• InceptionV4, Bi-LSTM, and LSTM layers are combined uniquely for the extraction of high-level features.
• The first algorithm for genre classification of Bollywood movies related to the Indian cinema.
• A novel EmoGDB dataset of 100 Bollywood movie trailers of six popular genres are developed.
• A novel idea in the field of affect-based video classification is devised for genre classification.
This dataset contains 100 Bollywood movie trailers in six popular and distinct genres: Action, Comedy, Drama, Horror, Romance, Thriller. The entire dataset is labeled with six induced emotions: Anger, Fear, Happy, Neutral, Sad, Surprise corresponding to every movie genre. This dataset can be used for movie genre classification and emotion classification. You can request this dataset by clicking on the following links:
https://drive.google.com/file/d/1TQrbGACWROBcrsA9Juw0HcOzdnq43kfo/view?usp=sharing https://drive.google.com/file/d/1iLUb2GZJgFFaRNKP371ASQoyKnuLKjYN/view?usp=sharing
This work is based on following paper. Please cite this paper:
Ashima Yadav and Dinesh Kumar Vishwakarma. "A unified framework of deep networks for genre classification using movie trailer." Applied Soft Computing 96 (2020): 106624 https://www.sciencedirect.com/science/article/pii/S1568494620305627
The authors would like to thank the following people for their sincere contribution in developing the EmoGDB dataset. The details of the contributors are:
1. Peya Monwar:: B.Tech in IT (Pursuing) from Delhi Technological University.
Her research area includes: Machine Learning, Sentimental Analysis.
LinkedIn : linkedin.com/in/jmpeya29/
Research Gate : www.researchgate.net/profile/Peya_Mowar
2. Mini Jain: B.Tech in IT (Pursuing) from Delhi Technological University.
Her research area includes: Machine Learning, Sentimental Analysis.
LinkedIn : linkedin.com/in/mini-jain-b5022a172/
Research Gate : https://www.researchgate.net/profile/Mini_Jain