AI-everydayy/HCMAI_2024_Batch-02_new_Keyframes

🎥 HCM AI Challenge 2024: Video Text Retrieval Dataset 🎯

🌟 Welcome to the HCMAI Challenge 2024!

This dataset has been prepared for the HCM AI Challenge 2024, designed specifically for video text retrieval tasks. The keyframes in this dataset have been extracted from various videos using the TransnetV2 model, providing highly accurate segmentation of scene transitions. The current batch, labeled Batch 02, includes keyframes from L13 to L24.

Certainly! Here's the updated README section with the DATASET_STATS.md cited as a link:

📊 Dataset Overview

The dataset contains keyframes extracted from videos across 12 folders (L13 - L24). Here's a summary of the video count per folder:

Batch 02: L13 - L24

• L13: 30 videos
• L14: 27 videos
• L15: 31 videos
• L16: 31 videos
• L17: 28 videos
• L18: 29 videos
• L19: 30 videos
• L20: 29 videos
• L21: 29 videos
• L22: 31 videos
• L23: 25 videos
• L24: 43 videos

Total videos: 363

For detailed statistics on each folder, including the number of keyframes, folder sizes, and individual video keyframe counts, please refer to the DATASET_STATS.md file. This file provides comprehensive information about the dataset structure and content.

Each video folder contains extracted keyframes (JPEG images) and a JSON file with frame timestamps. The number of keyframes varies between videos and folders, with detailed breakdowns available in the DATASET_STATS.md file.

🗂️ Folder Structure

The dataset is structured as follows:

./
├── assets
│   └── AIC2024-Banner-website.png
├── L13.zip
├── L14.zip
├── L15.zip
├── L16.zip
├── L17.zip
├── L18.zip
├── L19.zip
├── L20.zip
├── L21.zip
├── L22.zip
├── L23.zip
├── L24.zip
└── README.md

When extracted, each zip contains keyframes and corresponding metadata:

./
├── 📁 L13
│   ├── 📁 L13_V001
│   │   ├── 📄 L13_V001_frames_stamp.json
│   │   └── 📁 L13_V001_keyframes
│   │       ├── 🖼️ frame_0.jpg
│   │       ├── 🖼️ frame_1014.jpg
│   │       ├── 🖼️ frame_10128.jpg
│   │       └── ... (more frames)
│   ├── 📁 L13_V002
│   │   ├── 📄 L13_V002_frames_stamp.json
│   │   └── 📁 L13_V002_keyframes
│   │       ├── 🖼️ frame_0.jpg
│   │       └── ... (more frames)
│   └── ... (more videos)
└── ... (more lists from L14 to L24)

Each folder (L13 to L24) contains multiple videos, and each video folder consists of keyframes and a JSON file recording the timestamp for each frame.

🚀 Getting Started

To participate in the HCM AI Challenge 2024 🏆:

1. 📥 Download the dataset from the provided source or follow the steps below:

   1. 🔧 Setup Git LFS:

      curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash
      sudo apt-get install git-lfs
      

   2. 📂 Create and navigate to the Dataset directory:

      mkdir HCM_AI_Video_Text_Retrieval_Challenge_2024
      cd ./HCM_AI_Video_Text_Retrieval_Challenge_2024
      

   3. 🔗 Initialize Git and add the repository:

      git init
      git remote add origin https://huggingface.co/datasets/mtoan65/HCM_AI_Video_Text_Retrieval_Challenge_2024
      

   4. ⚙️ Install Git LFS hook for the repository:

      git lfs install
      

   5. ⬇️ Pull the repository:

      git checkout -b main
      git pull origin main
      

2. 🔍 Explore the dataset by unzipping the keyframe files in the Raw/Batch_02 directory.

📄 License

This dataset is distributed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license.

---

👤 About Me