Skip to content

kookmin-sw/2019-cap1-2019_3

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

유감(유튜브 감성분석)

팀페이지 주소 -> https://kookmin-sw.github.io/2019-cap1-2019_3

1. 프로잭트 소개

남녀노소 크리에이터가 기하급수적으로 증가하여 주목받고 있는데, 이들은 발 빠르게 트렌드를 읽어 시청자를 사로잡기 위해 고군분투 합니다. 크리에이터의 수만큼 각종 영상 콘텐츠들이 범람하는 지금, 그것을 평가하는 수치는 다양합니다. 하지만 콘텐츠의 평가를 가장 빠르게 보여주는 것은 댓글인데 영상에 달린 수많은 댓글을 모두 보는 것은 힘들어 전체 개관을 파악하고, 자신이 원하는 내용을 선택하여 읽기가 힘듭니다. 또한 시청자 반응의 지표를 볼 수 있는것은 댓글과 좋아요 시스템 뿐이라 크리에이터가 시청자의 반응을 파악하기에는 다소 부족합니다. 시청자의 입장에서는 수많은 영상들 중 본인이 원하는 영상만 골라보기 어렵습니다. 예를 들어, 전반적으로 웃긴 영상만 보고싶고 혐오적인 영상은 보고싶지 않은 경우 웃긴 영상만 골라서 찾아보기 힘듭니다. 또 시청자도 부정적인 악플은 보고싶지 않고 긍정적인 댓글만 골라서 보고싶은경우가 있지만 현재 유튜브에서 제공하는 댓글 필터는 없습니다. 따라서 본 프로젝트에서는 유튜브 영상을 분석하여 시청자에게는 원하는 감성의 영상과 댓글을 골라 볼 수 있도록 보여주고 크리에이터에게는 시청자의 반응을 웹캠으로 가져와서 감성을 분석한 결과를 제공하여 시청자가 어떤 감성으로 본인의 영상을 감상했는지 판단할 수 있는 새로운 지표를 제안함으로써 유튜브 크리에이터와 시청자가 양질의 영상을 창작하고 시청할 수 있도록 하기 위해 유감, 즉 유튜브에 대한 감성분석을 하는 프로젝트를 기획했습니다.

  • 크리에이터의 표정 변화를 분석해 타임라인을 제공합니다.
  • 유튜브 댓글 감성을 분석해 긍정, 중립, 부정 댓글을 분류합니다.
  • 6가지 감정인 행복, 슬픔, 화남, 놀람, 중립, 두려움으로 댓글의 감성 분포도를 제공합니다.
  • 시청자의 웹캠으로 받은 표정을 화남, 두려움, 경멸, 놀람, 행복, 슬픔, 중립으로 분석하여 시청자가 어떤 감성으로 동영상을 시청했는지 라인 차트로 나타냅니다.
  • 2. 소개 영상

    이 이미지를 클릭하시면 영상을 보실 수 있습니다. video

    3. 팀 소개

    OhTube

    정지연

    * email: 97jjy@kookmin.ac.kr
    * role: 팀장 ,댓글 감성 분석 , 댓글 전처리

    김주향

    * email: jh15501551@gmail.com
    * role: 영상 감성 분석  , 데이터 시각화

    손태선

    * email: escs34@gmail.com
    * role: 영상 감성 분석  , 영상 전처리

    이진주

    * role: 댓글 감성 분석  , 유튜브 댓글 크롤링

    양동철

    * email: moxenorljr@gmail.com
    * role: Web Front End 개발

    4.Abstract

    Creators are gaining attention as they grow exponentially, and they struggle to capture viewers by reading trends quickly. At a time when video content is flooded by the number of creators, there are various numbers to evaluate. However, the fastest way to evaluate content is to read comments, but it's hard to see all the comments in the video, so it's hard to see the whole view and choose what you want to read. And it's only a comment and a 'good' system that allows creators to see the viewer response indicators, so it's a little short of understanding the viewer's response. From the viewer's perspective, it's hard to select only the video you want out of the countless videos. For example, if you want to see funny images, but don't want to see hateful images, it's hard to find funny ones. In addition, viewers do not want to see negative comments, but only select positive comments, but there are no comment filters provided on YouTube. The project analyzes YouTube videos and shows viewers the choice of videos and comments they want. It also brings viewers' responses to the webcam and provides emotional analysis to creators. To help YouTube creators and viewers create and watch high-quality video, we've organized a project to analyze the emotions of YouTube, or YS.

  • Analyzes the changes in the Creator's facial expressions and provides a timeline.
  • Analyze the sensitivity of YouTube comments to categorize positive, neutral, and negative comments.
  • provide the emotional distribution of comments with six emotions: happiness, sadness, anger, surprise, neutrality, and fear.
  • Analyze the facial expressions received from your webcam with anger, fear, contempt, surprise, happiness, sadness, and neutrality to indicate in the line chart how the audience watched the video.
  • 5.시연영상

    video

    이미지를 클릭하시면 영상을 보실 수 있습니다.