PREDICTING EMOTION FROM COLOR PRESENT IN IMAGES AND VIDEO

Abstract

The goal of this project is to use machine learning techniques to predict behavior based on colors found in images and video excerpts. The aim of this paper is threefold: (a) to develop a machine learning algorithm to classify emotions according to the colors present in the image, (b) to select the best algorithm from the first stage and apply it to video excerpts as the color of emotional color. analyzing and (c) creating an online survey to check the accuracy of the transcribed video. First, three color extraction methods were tested, which were mixing the color into the color of the specified color, assigning the color to the RYB (red, yellow, blue) pattern, and extracting the color histograms contained in the residual image. This is based on the image's information, which contains color and emotional descriptions. Use multiple algorithms for classification, including deep learning and artificial intelligence techniques. Then, the results are obtained in the undifferentiated variable and the training set is given. In the second part, the best performing algorithm from the first stage was used to color the video excerpts according to emotions. This is followed thirdly by an online survey designed to examine the accuracy of respondents' algorithmic descriptions of film material. Additionally, the results obtained were discussed. The conclusion section contains a summary of the results and additional recommendations for improving the performance of the developped algorithm.