Exploiting character networks for movie summarization

Tran, Quang Dieu; Hwang, Dosam; Lee, O-Joun; Jung, Jai E.

doi:10.1007/s11042-016-3633-6

Cited by 32 publications

(14 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Basically, we expect that the protagonist will be likely to appear in each key scene within the story. Tran et al proposed the co-occurrence character network (CoCharNet) to represent the relationships between characters, as well as an approach to detect the protagonist within a movie by measuring the centrality and weighted degree of each node within the network graph [20]. In our work, we follow their definitions to build one character network for each episode and to extract story-related features from the network.…”

Section: Story-related Features: the Character Networkmentioning

confidence: 99%

“…The types of video studied by the community varies from self-made daily life amateur videos (e.g., from YouTube) through professional sports game recordings, to surveillance videos. Movies, as typical narrative videos, have been widely studied in order to address different challenges, such as movie summarization [20,30], movie description generation [31,32], and movie scene detection [33]. Current video understanding techniques excel at the clip level (a few seconds long videos) and, unlike our work, are not yet directed to narrated videos such as movies and TV series.…”

Section: Related Workmentioning

confidence: 99%

“…Tran et al proposed the idea of building a co-occurrence character network [20], which has been extended to an affective character network, representing the emotional relationship among characters. Based on this, Lee and Jung proposed a graph-style story model which reflects the story development [8].…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Towards story-based classification of movie scenes

Liu

Shmilovici

2020

PLoS ONE

View full text Add to dashboard Cite

Humans are entertained and emotionally captivated by a good story. Artworks, such as operas, theatre plays, movies, TV series, cartoons, etc., contain implicit stories, which are conveyed visually (e.g., through scenes) and audially (e.g., via music and speech). Story theorists have explored the structure of various artworks and identified forms and paradigms that are common to most well-written stories. Further, typical story structures have been formalized in different ways and used by professional screenwriters as guidelines. Currently, computers cannot yet identify such a latent narrative structure of a movie story. Therefore, in this work, we raise the novel challenge of understanding and formulating the movie story structure and introduce the first ever story-based labeled dataset-the Flintstones Scene Dataset (FSD). The dataset consists of 1, 569 scenes taken from a manual annotation of 60 episodes of a famous cartoon series, The Flintstones, by 105 distinct annotators. The various labels assigned to each scene by different annotators are summarized by a probability vector over 10 possible story elements representing the function of each scene in the advancement of the story, such as the Climax of Act One or the Midpoint. These elements are learned from guidelines for professional script-writing. The annotated dataset is used to investigate the effectiveness of various story-related features and multi-label classification algorithms for the task of predicting the probability distribution of scene labels. We use cosine similarity and KL divergence to measure the quality of predicted distributions. The best approaches demonstrated 0.81 average similarity and 0.67 KL divergence between the predicted label vectors and the ground truth vectors based on the manual annotations. These results demonstrate the ability of machine learning approaches to detect the narrative structure in movies, which could lead to the development of story-related video analytics tools, such as automatic video summarization and recommendation systems.

show abstract

Section: Story-related Features: the Character Networkmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Towards story-based classification of movie scenes

Liu

Shmilovici

2020

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Various studies have been conducted for extracting social networks in narrative artworks and discovering narrative features from the social networks [46]. These studies extracted the social networks from texts [47,48], videos [33,49,50], or both of them [51,52]. A few studies applied sentiment analysis to enrich the social networks [34,53].…”

Section: Related Workmentioning

confidence: 99%

Discovering Social Desires and Conflicts from Subculture Narrative Multimedia

et al. 2020

Self Cite

View full text Add to dashboard Cite

This study aims at discovering social desires and conflicts from subculture narrative multimedia. Since one of the primary purposes in the subculture consumption is vicarious satisfaction, the subculture works straightforwardly describe what their readers want to achieve and break down. The latent desires and conflicts are useful for understanding our society and realizing smart governance. To discover the social issues, we concentrate on that each subculture genre has a unique imaginary world that consists of inventive subjects. We suppose that the subjects correspond to individual social issues. For example, game fiction, one of the popular genres, describes a world like video games. Under game systems, everyone gets the same results for the same efforts, and it can be interpreted as critics for the social inequality issue. Therefore, we first extract subjects of genres and measure the membership degrees of subculture works for each genre. Using the subjects and membership degrees, we build a genealogy tree of subculture genres by tracing their evolution and differentiation. Then, we extract social issues by searching for the subjects that come from the real world, not imaginary. If a subculture work criticizes authoritarianism, it might include subjects such as government officials and bureaucrats. A combination of the social issues and genre genealogy tree will show diachronic changes in our society. We have evaluated the proposed methods by extracting social issues reflected in Korean web novels.

show abstract

“…Various studies [1][2][3][4] have been conducted for character networks (i.e., social networks between characters that appear in stories) to analyze stories in narrative multimedia (i.e., creative works that contain stories and are distributed through multimedia) automatically. They applied the analysis results on various applications-summarizing [5,6], recommending [7,8], indexing [9,10], and even generating [11] narrative multimedia. However, the character network model also has limitations.…”

Section: Introductionmentioning

confidence: 99%

Learning Hierarchical Representations of Stories by Using Multi-Layered Structures in Narrative Multimedia

Lee

Jung

Kim

2020

Sensors

Self Cite

View full text Add to dashboard Cite

Narrative works (e.g., novels and movies) consist of various utterances (e.g., scenes and episodes) with multi-layered structures. However, the existing studies aimed to embed only stories in a narrative work. By covering other granularity levels, we can easily compare narrative utterances that are coarser (e.g., movie series) or finer (e.g., scenes) than a narrative work. We apply the multi-layered structures on learning hierarchical representations of the narrative utterances. To represent coarser utterances, we consider adjacency and appearance of finer utterances in the coarser ones. For the movies, we suppose a four-layered structure (character roles ∈ characters ∈ scenes ∈ movies) and propose three learning methods bridging the layers: Char2Vec, Scene2Vec, and Hierarchical Story2Vec. Char2Vec represents a character by using dynamic changes in the character’s roles. To find the character roles, we use substructures of character networks (i.e., dynamic social networks of characters). A scene describes an event. Interactions between characters in the scene are designed to describe the event. Scene2Vec learns representations of a scene from interactions between characters in the scene. A story is a series of events. Meanings of the story are affected by order of the events as well as their content. Hierarchical Story2Vec uses sequential order of scenes to represent stories. The proposed model has been evaluated by estimating the similarity between narrative utterances in real movies.

show abstract

Exploiting character networks for movie summarization

Cited by 32 publications

References 25 publications

Towards story-based classification of movie scenes

Towards story-based classification of movie scenes

Discovering Social Desires and Conflicts from Subculture Narrative Multimedia

Learning Hierarchical Representations of Stories by Using Multi-Layered Structures in Narrative Multimedia

Contact Info

Product

Resources

About