Video2Music uses the deep-learning algorithm called Transformer to convert video/music input into mutually comparable latent vector features. By training a model using a large number of movie contents online*. Using the Contrastive Learning technique, we successfully calculate quantitatively the fitness between a video and a song, two distinct forms of media. The pre-trained model provided with the product license already supports a wide range of input videos and music styles but can be re-trained with additional data to improve accuracy for specific applications.

*Training machine learning models from copyrighted materials is permitted by the copyright law of Japan