Serial Speakers: a Dataset of TV Series

Dataset of three TV Series with manual annotations. Cite as:@inproceedings{Bost2020, title = {Serial Speakers: a Dataset of TV Series}, author = {Bost, Xavier and Labatut, Vincent and Linares, Georges}, url = {https://hal.archives-ouvertes.fr/hal-02477736}, booktitle = {12th International Conference on Language Resources and Evaluation (LREC 2020)}, address = {Marseille, France}, year = {2020}} The dataset consists of 3 TV series: - Breaking Bad: S01--S05 (file 'bb.json') - Game of Thrones: S01--08 (file 'got.json') - House of Cards: S01--S02 (file 'hoc.json') All three files are in .json format and contain TV Series annotated data. Each TV Series is defined by its name, A TV Series contains seasons, defined by their ids. Every season is made of episodes, defined by their ids, titles, duration and fps . Each episode contains two basic kinds of data: scenes and speech segments. Scenes are defined by starting points and are made of shots (Seasons 1 only). A shot is defined by: - Starting and ending positions.- Recurring shot ids. The speech segments are defined by their: - Starting and ending points. - Textual content (here encrypted for copyright reasons). - Speaker. - Possible interlocutors (for the following episodes only: bb: S01E04, S01E06, S02E03, S02E04; got: S01E03, S01E07, S01E08; hoc: S01E01, S01E07, S01E11). All timestamps are expressed in seconds and are valid for the video files extracted from the commercial DVDs (PAL 25 FPS), with recaps (unannotated) included at the beginning of the House of Cards episodes. In you are interested in the textual content of the dataset, please consider using our text recovering tool on GitHub: https://github.com/bostxavier/Serial-Speakers A comprehensive description of the dataset can be found at: https://hal.archives-ouvertes.fr/hal-02477736本数据集收录三部带人工标注的电视剧标注数据，引用格式如下：@inproceedings{Bost2020, title = {系列演讲者：电视剧数据集}, author = {Bost, Xavier and Labatut, Vincent and Linares, Georges}, url = {https://hal.archives-ouvertes.fr/hal-02477736}, booktitle = {第12届国际语言资源与评估会议（LREC 2020）}, address = {法国马赛}, year = {2020}} 本数据集包含三部电视剧： - 《绝命毒师》（Breaking Bad）：第1季至第5季，对应文件`bb.json` - 《权力的游戏》（Game of Thrones）：第1季至第8季，对应文件`got.json` - 《纸牌屋》（House of Cards）：第1季至第2季，对应文件`hoc.json` 所有文件均采用JSON格式（JavaScript Object Notation），存储对应电视剧的标注数据。每部电视剧以名称标识，包含若干季（以季ID定义）；每一季由若干集构成，集以ID、标题、时长及帧率（fps）定义。每一集包含两类基础数据：场景（scene）与语音片段（speech segments）。 场景以起始点作为标识，由镜头（shot，仅第1季包含）构成。镜头的定义包括： - 起始与结束位置 - 循环镜头编号 语音片段的定义包括： - 起始与结束时间点 - 文本内容（因版权原因已加密） - 说话者 - 潜在对话者（仅适用于以下剧集：《绝命毒师》：S01E04、S01E06、S02E03、S02E04；《权力的游戏》：S01E03、S01E07、S01E08；《纸牌屋》：S01E01、S01E07、S01E11） 所有时间戳均以秒为单位，适用于从商业DVD提取的视频文件（PAL制式25 FPS），且《纸牌屋》剧集开头包含的回顾片段（未标注）已被纳入标注范围。 若您需要获取本数据集的原始文本内容，可使用我们在GitHub上开源的文本恢复工具：https://github.com/bostxavier/Serial-Speakers 本数据集的详细描述可参阅：https://hal.archives-ouvertes.fr/hal-02477736