SpeechPT

프로메테우스 AI 발표연습 어플리케이션 개발

Table of Contents

• General info
• Technologies
• Setup
• Members

How to Use

1. Clone the repository :

git clone https://github.com/use08174/SpeechPT.git

2. Install dependencies:

pip install -r requirements.txt

3. Start the development server:

cd SpeechPT/backend/app
uvicorn main:app --reload

Key Features

Our project has two key features: presentation analysis and text summarization.

Presentation analysis

Speech Analysis

After STT using Google STT API, we implemented three functions as follow.

• Speed : Calculate cpm to analyze speed of the speech. (cpm : character per minute)
• Pause : Calculate number of pauses and duration of pauses during speech.
• Filler words : Detect korean filler words in the speech.

Video Analysis

• Face Expression : Analyze face expression of the presenter.
• Gaze Tracking : Tracking gaze of the presenter.

Text summarization

• Summarize text
  1. Extract text from a document (pdf,word) and summarize text by paragraph using Kobart
  2. Check the spelling and save the final summary to the document

Our Project Archive

This is our notion page for our project archive. : Notion

Members

멤버이름	역할
심수민	Leader, Audio, Backend
배아람	Audio, Frontend
김수홍	Audio, Frontend
문승현	Video, Backend
남승현	Video, Backend
윤선미	Text, Frontend
임서연	Text, Backend
안서연	UI Design

More about Prometheus : https://prometheus-ai.net/