On the influence of low-level visual features in film classification

Abstract

Background In this paper we present a model of parameters to aesthetically characterize films using a multi-disciplinary approach: by combining film theory, visual low-level video descriptors (modeled in order to supply aesthetic information) and classification techniques using machine and deep learning. Methods Four different tests have been developed, each for a different application, proving the model's usefulness. These applications are: aesthetic style clustering, prediction of production year, genre detection and influence on film popularity. Results The results are compared against high-level information to determine the accuracy of the model to classify films without knowing such information previously. The main difference with other film characterization approaches is that we are able to isolate the influence of high-level descriptors to really understand the relevance of low-level features and, accordingly propose a useful set of low-level visual descriptors for that purpose. This model has been tested with a representative number of films to prove that it can be used for different applications.

Poster-Based Multiple Movie Genre Classification Using Inter-Channel Features

“…Low-level visual characteristics, such as the colors and edges of the poster, were extracted using a color histogram and a GIST image descriptor [7], [8]. Genre classification was performed using conventional machine learning techniques [11], [12]. Multi-label k-nearest neighbor and Random K-labelsets algorithms were used for the classification of multiple label data; single label data were also classified using the Naive Bayes classifier.…”

“…With the use of deep learning in genre classification, the number of posters used gradually increased [10], [13], [14]. However, many studies used both conventional machine learning techniques and statistical multi-label methods, rather than using only deep learning models [11]. Chu and Guo [1] performed genre classification beyond the existing machine learning techniques by applying a convolution network to the poster images.…”

“…Similar to genre classification using posters, the use of lowlevel features for classification in movie trailers has also been studied [11], [15]. In addition, genre classification based on trailer information has been performed using convolutional neural networks [16].…”

Exploration of the Characteristics of Emotion Distribution in Korean TV Series: Common Pattern and Statistical Complexity

“…Cutting [47] used features such as shot transition, motion, shot type, and shot duration to observe the narrative dynamics of popular movies. Álvarez et al [48] used 24 features in three categories-image, pace, and motion-to predict the genre, production year, and popularity of movies.…”

“…In the literature, there has been recent interest in the complexity of media content in various forms. Image complexity has been measured by a variety of metrics, including fractal dimension [58], algorithmic specified complexity [59], compressed file size [60], the degree of flatness in the profile of gray-scale mean variance [61], and the entropy of luminosity [48]. In addition, the complexities of text discourse [62], movie narration [63], and the rhythm of films [64] have been investigated.…”