🤖📊

synslidegen

Generate copyright-free synthetic slides with ease!
[Website] [Paper] [Dataset]

· Report Bug · Request Feature

Table of Contents

1. About The Project
2. Getting Started
   • Prerequisites
   • Installation
3. Usage
4. Roadmap
5. Contact
6. Acknowledgments

About The Project

synslidegen is a modular and lightweight pipeline to generate synthetic slides with automated annotations for Slide Element Detection and Text Query-based Slide Retrieval tasks. Open-sourced code as part of "SynSlideGen : AI-Generated Lecture Slides for Improving Slide Element Detection and Retrieval" paper to be presented in ICDAR 2025

(back to top)

SynDet(top) and SynRet(bottom) samples

Getting Started

Follow these instructions to run the tool locally on your system.

Prerequisites

• Python v3.6>
• pip v22.1>
• Microsoft Powerpoint Viewer
• MS-Office Supported OS (Windows 7 or newer)

Installation

1. Clone the repo

   git clone https://github.com/synslidegen/synslidegen_pipeline

2. Create a new Python virtual environment

   python3 -m venv <virtual-environment-name>

3. Go the root folder of the project and install dependencies using pip

   pip install -r requirements.txt

(back to top)

Usage

To modify the topic and content of the presentation slides

Go to the code\config.py file and modify the SUBJECT and BOOK variables as well as other parameters like MAX_TOPIC, LLM_MODEL, TEMPERATURE, etc. You can also modify the DUP_FAC to generate more presentation from a single topic and TEMPLATE_PERCENTAGE to control the percentage of design-template slides in the generations.

To generate synthetic presentation slides

In your Powershell prompt, go the project root folder and run the following script

 .\run_synslidegen.ps1 -py_path "\Custom\Python\Path" -env_path "\Custom\VirtualEnv\Path"

(Make sure you have Set-Execution-Policy: RemoteSigned)

You will find your presentation slides saved in the folder

 ppts\

To generate automated annotations for Slide Element Detection

In your Powershell prompt, go the project root folder and uncomment the following lines in the run_synslidegen.ps1 file

 #python 'D:\Research_work\Experimentation_Results\py_pptx_code_gen_using_LLMs\pptGEN_CLEANED\code\post_processing\pptx_to_png.py'
 #python 'D:\Research_work\Experimentation_Results\py_pptx_code_gen_using_LLMs\pptGEN_CLEANED\code\post_processing\bbox_annotations.py'

(Make sure you have Set-Execution-Policy: RemoteSigned)

To generate automated annotations for Slide Image Retrieval

In your Powershell prompt, go the project root folder and uncomment the following lines in the run_synslidegen.ps1 file

 #python 'D:\Research_work\Experimentation_Results\py_pptx_code_gen_using_LLMs\pptGEN_CLEANED\code\post_processing\summary_gen_from_json.py'
 #python 'D:\Research_work\Experimentation_Results\py_pptx_code_gen_using_LLMs\pptGEN_CLEANED\code\post_processing\rephrase_summary.py'

(Make sure you have Set-Execution-Policy: RemoteSigned)

(back to top)

Roadmap

• Add v0.1 code
• Provide Demo Images of generated PPTs
• Execute code pipeline using bash script
• Enable content coherence
• Make the tool web-accessible

See the open issues for a full list of proposed features (and known issues).

(back to top)

Citation

@inproceedings{synslidegen,
  title={SynSlideGen: AI-Generated Lecture Slides for Improving Slide Element Detection and Retrieval},
  author={Suyash Maniyar, Vishvesh Trivedi, Ajoy Mondal, Anand Mishra, C.V. Jawahar},
  booktitle={ICDAR},
  year={2025}
}