遇见数据集

Serial Speakers: a Dataset of TV Series

收藏
Figshare2020-02-17 更新2026-04-08 收录
官方服务:

资源简介:

<b>Dataset of three TV Series</b> with <b>manual</b> annotations.<br>Cite as:@inproceedings{Bost2020,<br> title = {Serial Speakers: a Dataset of TV Series},<br> author = {Bost, Xavier and Labatut, Vincent and Linares, Georges},<br> url = {https://hal.archives-ouvertes.fr/hal-02477736},<br> booktitle = {12th International Conference on Language Resources and Evaluation (LREC 2020)},<br> address = {Marseille, France},<br> year = {2020}} <pre><br>The dataset consists of 3 TV series: <br></pre><br>- <i>Breaking Bad</i>: S01--S05 (file 'bb.json')<br>- <i>Game of Thrones</i>: S01--08 (file 'got.json')<br>- <i>House of Cards</i>: S01--S02 (file 'hoc.json')<br><br>All three files are in .json format and contain TV Series annotated data.<br><br>Each TV Series is defined by its <b>name</b>,<br><br>A TV Series contains <b>seasons</b>, defined by their <b>id</b>s.<br><br>Every season is made of <b>episodes</b>, defined by their <b><b>id</b></b>s,<b> title</b>s, <b>duration </b>and<b> fps </b>.<br><br>Each episode contains two basic kinds of <b>data</b>: <b>scenes</b> and <b>speech segments</b>.<br><br>Scenes are defined by <b>start</b>ing points and are made of <b>shots </b>(Seasons 1 only)<b>.<br><br></b>A shot is defined by<b>:<br><br></b>- <b>Start</b>ing and <b>end</b>ing positions.-<b> </b>Recurring shot <b>id</b>s.<br>The speech segments are defined by their:<br>- <b>Start</b>ing and <b>end</b>ing points.<br>- <b>Text</b>ual content (here encrypted for copyright reasons).<br>- <b>Speaker</b>.<br>- Possible<b> interlocutors</b> (for the following episodes only: <b>bb</b>: S01E04, S01E06, S02E03, S02E04; <b>got</b>: S01E03, S01E07, S01E08; <b>hoc</b>: S01E01, S01E07, S01E11).<br><br>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 <i>House of Cards</i> episodes.<br>In you are interested in the textual content of the dataset, please consider using our text recovering tool on GitHub:<br>https://github.com/bostxavier/Serial-Speakers<br>A comprehensive description of the dataset can be found at:<br>https://hal.archives-ouvertes.fr/hal-02477736<br><b> </b><br>

**三部电视剧数据集**(带**人工标注**)。 引用格式如下: @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} } 本数据集包含三部电视剧: - *《绝命毒师》(Breaking Bad)*:第1至5季(对应文件`bb.json`) - *《权力的游戏》(Game of Thrones)*:第1至8季(对应文件`got.json`) - *《纸牌屋》(House of Cards)*:第1至2季(对应文件`hoc.json`) 上述三个文件均为`.json`格式,存储电视剧标注数据。 每部电视剧以其**名称**作为标识。 电视剧包含**季(season)**,以**编号(id)**进行区分。 每一季由若干**集(episode)**组成,集以**编号(id)**、**标题(title)**、**时长(duration)**和**帧率(fps)**作为标识。 每一集包含两类基础**数据**:**场景(scene)**与**语音片段(speech segment)**。 场景以**起始**时间点作为标识,由**镜头(shot)**组成(仅第一季包含镜头信息)。 镜头的定义包括: - **起始**与**终止**时间位置 - 重复使用的镜头**编号(id)** 语音片段的定义包括: - **起始**与**终止**时间点 - **文本**内容(因版权原因已加密) - **发言者(speaker)** - 可选**对话对象(interlocutor)**,仅适用于以下集数: - 《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帧率);《纸牌屋》剧集开头包含的回顾片段未进行标注。 若您需要获取本数据集的**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 本数据集的详细说明可访问:https://hal.archives-ouvertes.fr/hal-02477736

创建时间:
2020-02-08
二维码
社区交流群
二维码
科研交流群
商业服务