index.html

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <title>Enhancing 3D Point Cloud Classification</title>
  <style>
    /* Reset and base styles */
    * { margin: 0; padding: 0; box-sizing: border-box; }
    body { font-family: 'Helvetica Neue', Arial, sans-serif; line-height: 1.6; color: #333; background: #f5f5f5; }
    
    header { 
      background: #1e1e1e; 
      color: #fff; 
      padding: 20px 0; 
      text-align: center; 
      position: sticky; 
      top: 0; 
      z-index: 1000;
    }
    header h1 { margin-bottom: 10px; font-size: 2.5em; }
    nav ul { list-style: none; display: flex; justify-content: center; flex-wrap: wrap; }
    nav ul li { margin: 0 15px; }
    nav ul li a { color: #fff; text-decoration: none; font-size: 1.1em; transition: color 0.3s; }
    nav ul li a:hover { color: #007acc; }

    /* Sources Section */
    .sources {
      background: #fff;
      padding: 10px 0;
      text-align: center;
      box-shadow: 0 2px 4px rgba(0,0,0,0.1);
    }
    .sources a {
      margin: 0 10px;
      text-decoration: none;
    }
    
    /* Hero Section */
    .hero { 
      position: relative; 
      background: url('https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/Point-SkipNet_Overview.jpg?raw=true') no-repeat center center;
      background-size: contain; 
      height: 60vh; 
      display: flex; 
      align-items: center; 
      justify-content: center; 
      text-align: center; 
      padding: 0 20px; 
    }
    .hero::before { 
      content: ""; 
      position: absolute; 
      top: 0; left: 0; 
      width: 100%; height: 100%; 
      background-color: rgba(0, 0, 0, 0.5); 
      z-index: 1; 
    }
    .hero > div { 
      position: relative; 
      z-index: 2; 
      color: #fff; 
    }
    .hero h2 { font-size: 3em; margin-bottom: 10px; }
    .hero p { font-size: 1.3em; }
    
    /* Sections */
    section { 
      padding: 60px 20px; 
      max-width: 1000px; 
      margin: 20px auto; 
      background: #fff; 
      border-radius: 8px; 
      box-shadow: 0 2px 8px rgba(0,0,0,0.1); 
    }
    section h2 { 
      text-align: center; 
      margin-bottom: 40px; 
      font-size: 2em; 
      border-bottom: 2px solid #eee; 
      padding-bottom: 10px;
    }
    section p, section li { margin-bottom: 20px; font-size: 1.1em; }
    section pre { background: #f0f0f0; padding: 15px; border-radius: 5px; overflow-x: auto; }
    
    /* Two-column layout for Overview */
    .overview-container {
      display: flex; 
      align-items: center; 
      justify-content: space-between; 
      gap: 20px;
      flex-wrap: wrap;
    }
    .overview-text { flex: 1; min-width: 280px; }
    
    /* Figures */
    figure { text-align: center; margin: 30px 0; }
    figure img { max-width: 80%; height: auto; border-radius: 4px; }
    .small-image img { width: 50%; }
    figure figcaption { margin-top: 10px; font-size: 0.95em; color: #555; }
    
    /* Video Embed */
    .video-container {
      position: relative; 
      padding-bottom: 56.25%; 
      height: 0; 
      overflow: hidden; 
      max-width: 100%; 
      margin: 20px auto;
    }
    .video-container iframe {
      position: absolute; 
      top: 0; 
      left: 0; 
      width: 100%; 
      height: 100%;
    }
    
    /* Resource Cards */
    .resource-container {
      display: flex; 
      justify-content: space-around; 
      flex-wrap: wrap; 
      gap: 20px; 
      text-align: center; 
    }
    .resource-card {
      background: #fff; 
      border-radius: 8px; 
      box-shadow: 0 4px 8px rgba(0,0,0,0.1); 
      padding: 20px; 
      width: 300px; 
      transition: transform 0.3s;
    }
    .resource-card:hover { transform: translateY(-5px); }
    .resource-card img { width: 80px; margin-bottom: 10px; }
    .resource-card h3 { font-size: 1.4em; margin-bottom: 10px; }
    .resource-card a { text-decoration: none; color: #007acc; font-weight: bold; }
    .resource-card a:hover { text-decoration: underline; }
    
    footer { 
      text-align: center; 
      padding: 20px; 
      background: #1e1e1e; 
      color: #fff; 
      font-size: 0.9em; 
    }
    
    @media (max-width: 768px) {
      .overview-container { flex-direction: column; text-align: center; }
    }
  </style>
</head>
<body>
  <header>
    <h1>Enhancing 3D Point Cloud Classification</h1>
    <nav>
      <ul>
         <li><a href="#overview">Overview</a></li>
         <li><a href="#video">Video</a></li>
         <li><a href="#challenges">Challenges</a></li>
         <li><a href="#solution">Solution</a></li>
         <li><a href="#dataset">Dataset</a></li>
         <li><a href="#training">Training & Testing</a></li>
         <li><a href="#experiments">Experiments</a></li>
         <li><a href="#resources">Resources</a></li>
         <li><a href="#acknowledgements">Acknowledgements</a></li>
      </ul>
    </nav>
  </header>

  <!-- Sources Section -->
  <div class="sources">
    <a href="https://arxiv.org/link_paper" target="_blank">
      <img src="https://img.shields.io/badge/Paper-arXiv-brightgreen" alt="Paper on arXiv"/>
    </a>
    <a href="https://github.com/m-saeid/ModeNetR_PointSkipNet" target="_blank">
      <img src="https://img.shields.io/badge/Project-GitHub-red" alt="GitHub Repository"/>
    </a>
    <a href="https://www.youtube.com/watch?v=7ziipjpdth0&list=PLvWl5fdJgzQxaF0v4egv1cdrstl8N7fEM&index=2" target="_blank">
      <img src="https://img.shields.io/badge/Video-YouTube-blue" alt="YouTube Presentation"/>
    </a>
    <a href="https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/ModelNet%E2%80%91R%20%26%20Point%E2%80%91SkipNet.pdf" target="_blank">
      <img src="https://img.shields.io/badge/Presentation-PDF-orange" alt="Presentation PDF"/>
    </a>
  </div>

  <!-- Overview Section with Two-column Layout -->
  <section id="overview">
    <h2>Overview</h2>
    <div class="overview-container">
      <div class="overview-text">
        <p><strong>ModelNet‑R</strong> is our refined version of the ModelNet40 dataset. By resolving labeling inconsistencies, removing near‑2D data, and improving class differentiation, we provide a cleaner, more reliable benchmark for 3D point cloud classification.</p>
        <p><strong>Point‑SkipNet</strong> is a lightweight, graph‑based neural network designed to efficiently extract features from 3D point clouds using smart sampling, grouping, and skip connections. This combination achieves state‑of‑the‑art accuracy with lower computational overhead.</p>
      </div>
    </div>
  </section>

  <!-- Video Presentation Section -->
  <section id="video">
    <h2>Video Presentation</h2>
    <div class="video-container">
      <iframe src="https://www.youtube.com/embed/7ziipjpdth0" frameborder="0" allowfullscreen></iframe>
    </div>
  </section>

  <!-- Challenges Section -->
  <section id="challenges">
    <h2>Challenges in ModelNet40</h2>
    <ul>
      <li>Inconsistent labeling and ambiguous class definitions</li>
      <li>Presence of nearly 2D objects lacking volumetric depth</li>
      <li>Size mismatches causing classification confusion</li>
    </ul>
    <figure class="small-image">
      <img src="https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/ModelNet_Problems.jpg?raw=true" alt="Challenges in ModelNet40">
      <figcaption>Figure: Challenges in the ModelNet40 dataset</figcaption>
    </figure>
  </section>

  <!-- Our Solution Section -->
  <section id="solution">
    <h2>Our Solution: ModelNet‑R & Point‑SkipNet</h2>
    <ol>
      <li><strong>ModelNet‑R Dataset:</strong> A refined dataset featuring corrected labels, removal of 2D artifacts, and enhanced class differentiation.</li>
      <li><strong>Point‑SkipNet Architecture:</strong> A lightweight neural network that leverages efficient point sampling, grouping, and skip connections for improved feature extraction.</li>
    </ol>
    <figure>
      <img src="https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/Point-SkipNet.jpg?raw=true" alt="Point‑SkipNet Architecture">
      <figcaption>Figure: The Point‑SkipNet architecture</figcaption>
    </figure>
    <figure class="small-image">
      <img src="https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/SampleAndGroup.jpg?raw=true" alt="Sample and Group Module">
      <figcaption>Figure: Sample and Group Module for efficient point sampling</figcaption>
    </figure>
  </section>

  <!-- Dataset Section -->
  <section id="dataset">
    <h2>Dataset</h2>
    <p>Place these datasets in the <code>data</code> folder:</p>
    <ul>
      <li><code>data/modelnet40_normal_resampled</code></li>
      <li><code>data/modelnet40_ply_hdf5_2048</code></li>
    </ul>
    <p><em>Note:</em> The <code>modelnet40_normal_resampled</code> dataset will be converted to <code>modelnetR_normal_resampled</code> during processing. Please back up the original data if needed.</p>
  </section>

  <!-- Training & Testing Section -->
  <section id="training">
    <h2>Training &amp; Testing</h2>
    <p>To train Point‑SkipNet:</p>
    <pre><code>cd Point-SkipNet
dataset="modelnetR" bash run_train.sh</code></pre>
    <p>To test the model:</p>
    <pre><code>cd Point-SkipNet
dataset="modelnetR" bash run_test.sh</code></pre>
  </section>

  <!-- Experimental Results Section -->
  <section id="experiments">
    <h2>Experimental Results</h2>
    <p>Our experiments demonstrate significant improvements using ModelNet‑R and Point‑SkipNet:</p>
    <ul>
      <li><strong>ModelNet:</strong> Overall Accuracy ~92.29%, Mean Class Accuracy ~89.84%</li>
      <li><strong>ModelNet‑R:</strong> Overall Accuracy ~94.33%, Mean Class Accuracy ~92.93%</li>
    </ul>
  </section>

  <!-- Resources Section -->
  <section id="resources">
    <h2>Resources</h2>
    <div class="resource-container">
      <div class="resource-card">
        <img src="https://img.shields.io/badge/Paper-arXiv-brightgreen" alt="Paper on arXiv">
        <h3>Read the Paper</h3>
        <p>Explore the full methodology and results in our publication.</p>
        <a href="https://arxiv.org/link_paper" target="_blank">View Paper</a>
      </div>
      <div class="resource-card">
        <img src="https://img.shields.io/badge/Project-GitHub-red" alt="GitHub Repository">
        <h3>GitHub Repository</h3>
        <p>Access the complete code, dataset, and documentation.</p>
        <a href="https://github.com/m-saeid/ModeNetR_PointSkipNet" target="_blank">Visit GitHub</a>
      </div>
      <div class="resource-card">
        <img src="https://img.shields.io/badge/Video-YouTube-blue" alt="YouTube Presentation">
        <h3>YouTube Video</h3>
        <p>Watch our detailed presentation on YouTube.</p>
        <a href="https://www.youtube.com/watch?v=7ziipjpdth0&list=PLvWl5fdJgzQxaF0v4egv1cdrstl8N7fEM&index=2" target="_blank">Watch Video</a>
      </div>
      <div class="resource-card">
        <img src="https://img.shields.io/badge/Presentation-PDF-orange" alt="Presentation PDF">
        <h3>Download Presentation</h3>
        <p>Get the PDF presentation for an in-depth overview.</p>
        <a href="https://github.com/m-saeid/ModeNetR_PointSkipNet/blob/main/images/ModelNet%E2%80%91R%20%26%20Point%E2%80%91SkipNet.pdf" target="_blank">Download PDF</a>
      </div>
    </div>
  </section>

  <!-- Acknowledgements Section -->
  <section id="acknowledgements">
    <h2>Acknowledgements</h2>
    <p>This project builds upon the valuable contributions of the research community. Special thanks to:</p>
    <ul>
      <li><strong>PointNet++:</strong> Deep Hierarchical Feature Learning on Point Sets in a Metric Space (<a href="https://arxiv.org/pdf/1706.02413" target="_blank">arXiv</a>)</li>
      <li><strong>GitHub Implementation:</strong> <a href="https://github.com/yanx27/Pointnet_Pointnet2_pytorch" target="_blank">PointNet++ PyTorch</a></li>
      <li><strong>ModelNet Dataset:</strong> The original benchmark from 3D ShapeNets (<a href="https://openaccess.thecvf.com/content_cvpr_2015/papers/Wu_3D_ShapeNets_A_2015_CVPR_paper.pdf" target="_blank">PDF</a>)</li>
    </ul>
  </section>

  <footer>
    <p>© 2025 ModelNet-R &amp; Point-SkipNet Project.</p> <!-- Released under Apache 2.0 License.</p> -->
  </footer>
</body>
</html>