Mise-en-scene recommendation system

Data di pubblicazione




Data di priorità



US patent application


Politecnico di Milano


Department of Electronics, Information and Bioengineering


Cremonesi Paolo, Elahi Mehdi, Deldjoo Yashar


Recommendations are traditionally generated on the basis of user’s implicit and explicit preferences on movies’ attributes, such as genre, director, and actors. However, user’s preferences can be better described by the mise-en-sc`ene characteristics of movies namely the design aspects of a movie production used to classify aesthetic and style. Lighting, colors, background, and movements in a movie are all examples of mise-en-sc`ene features. Although viewers may not consciously notice movie style, it still affects the viewer’s experience of the movie. The mise-en-sc`ene highlights similarities in the narratives, as movie makers typically relate the overall movie style to reflect the story, and can be used to categorize movies at a finer level compared to the traditional movie features. The invention is a method and a system for personalized multimedia recommendation. The system analyzes the multimedia content and extracts the audio- visual features (attributes), grounded on the style of the movies, which form the Mise-en-sc`ene characteristics. The system uses the extracted information to generate relevant recommendations for users. This process allows the multi- media recommender system to deal with the Cold Start problem, which occurs when the system is unable to accurately recommend a new multimedia item to the users, since the new item is added to the catalogue and no information (e.g., genre, cast, producer, date of production, ratings, reviews, tags, and description) is available for the movie item. This is a situation that typically occurs in social movie-sharing web applications where every day, hundred millions of hours of videos are uploaded by users with no related information.

Campo di applicazione

Multimedia recommendation systems.


This technique: • is totally automatic and it does not necessarily require human involvement, in contrast to the similar techniques that require human or even expert involvement; • can be adopted in the cold start new item situation, that is when the other techniques fail to work properly; • is not computationally expensive and is completely scalable to big data, in contrast to the other techniques which require expensive process such as data pre-processing; • is bridge between artistic view toward movie making and the technical view and fills the semantic gap between the artists and engineers.

Stadio di sviluppo

Working algorithm tested on a wide dataset of movies.