Emotion recognition has always been one of the interesting topics for researchers in the field of Artificial Intelligence and Machine Learning, extending the findings towards human and computer interactions. Most of the existing research and state-of-the-art solutions are based on emotion analysis on datasets based on facial expressions, some of which include CelebA, Aff-Wild, Hateful Memes, Yale Face Database, etc. The former research based on facial expressions, body language, or even audio signals does not have a wide variety in the classification of emotions, which focuses on video and audio analysis.

In this paper, we present a single-modal, multi-labeled image dataset. Our approach focuses on creating a novel image dataset by extracting the images from some randomly selected videos from movies/series. This multi-labeled dataset depicts the emotional impact of the sequence of images from the video on the target viewer. Our proposed method and technique for creating the dataset is implemented using Python and the OpenCV library, with multi-label classification performed by raters and reviewers falling under the category of teenagers. The novel dataset would be used for performing an experiment to help us understand the emotional impact of the test data and study the impact of the generated sequence of emotions present in any video clip on any teenager.