| <!DOCTYPE html> |
| <html lang="en"> |
| <head> |
| <meta charset="UTF-8" /> |
| <meta name="viewport" content="width=device-width, initial-scale=1.0" /> |
| <title>4× Super-Resolution cGAN for Agricultural Leaf Images</title> |
| <script src="https://cdn.tailwindcss.com"></script> |
| <script src="https://cdn.jsdelivr.net/npm/chart.js"></script> |
| <link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;600;700;800&display=swap" rel="stylesheet"> |
| <style> |
| body { font-family: 'Inter', sans-serif; } |
| .gradient { background: linear-gradient(135deg, #052e16 0%, #15803d 50%, #22c55e 100%); } |
| .glass { backdrop-filter: blur(10px); background: rgba(255,255,255,0.78); } |
| .code-card { background: #0f172a; color: #e2e8f0; } |
| pre { white-space: pre-wrap; } |
| </style> |
| </head> |
| <body class="bg-slate-50 text-slate-800"> |
| <header class="gradient text-white"> |
| <div class="max-w-6xl mx-auto px-6 py-20"> |
| <p class="uppercase tracking-[0.3em] text-green-200 text-sm mb-4">Computer Vision Project</p> |
| <h1 class="text-5xl md:text-6xl font-extrabold leading-tight">4× Super-Resolution cGAN</h1> |
| <p class="mt-6 text-xl text-slate-100 max-w-3xl leading-8"> |
| A conditional GAN that reconstructs high-resolution agricultural leaf images from low-resolution inputs. |
| The model combines adversarial training, perceptual loss, and pixel-wise reconstruction loss to preserve |
| fine textures and disease patterns critical for plant health analysis. |
| </p> |
| <div class="mt-8 flex flex-wrap gap-3"> |
| <span class="px-4 py-2 rounded-full bg-white/10 border border-white/20">4× Upscaling</span> |
| <span class="px-4 py-2 rounded-full bg-white/10 border border-white/20">Conditional GAN</span> |
| <span class="px-4 py-2 rounded-full bg-white/10 border border-white/20">VGG Perceptual Loss</span> |
| <span class="px-4 py-2 rounded-full bg-white/10 border border-white/20">Agricultural Images</span> |
| </div> |
| </div> |
| </header> |
|
|
| <main class="max-w-6xl mx-auto px-6 py-12 space-y-20"> |
| <section class="grid md:grid-cols-4 gap-6 -mt-20"> |
| <div class="glass rounded-3xl shadow-xl p-6"><div class="text-3xl font-extrabold text-green-600">4×</div><div class="text-sm text-slate-600 mt-1">Resolution Enhancement</div></div> |
| <div class="glass rounded-3xl shadow-xl p-6"><div class="text-3xl font-extrabold text-emerald-600">cGAN</div><div class="text-sm text-slate-600 mt-1">Generator + Discriminator</div></div> |
| <div class="glass rounded-3xl shadow-xl p-6"><div class="text-3xl font-extrabold text-lime-600">VGG</div><div class="text-sm text-slate-600 mt-1">Perceptual Loss</div></div> |
| <div class="glass rounded-3xl shadow-xl p-6"><div class="text-3xl font-extrabold text-teal-600">MAE</div><div class="text-sm text-slate-600 mt-1">Optimization Target</div></div> |
| </section> |
|
|
| <section> |
| <h2 class="text-3xl font-bold mb-4">Project Overview</h2> |
| <p class="text-lg text-slate-600 leading-8"> |
| This project addresses image super-resolution for agricultural leaf datasets. Low-resolution images |
| are upscaled by a factor of four using a conditional GAN with a deep residual generator and a |
| multi-scale discriminator. A hybrid objective combines adversarial loss, L1 reconstruction loss, |
| and VGG-based perceptual loss to generate visually realistic and structurally accurate outputs. |
| </p> |
| </section> |
|
|
| <section> |
| <h2 class="text-3xl font-bold mb-8">Pipeline Architecture</h2> |
| <div class="grid md:grid-cols-5 gap-4 text-center"> |
| <div class="bg-white rounded-2xl p-5 shadow">Low-Res Input</div> |
| <div class="bg-white rounded-2xl p-5 shadow">Residual Generator</div> |
| <div class="bg-white rounded-2xl p-5 shadow">4× Upsampling</div> |
| <div class="bg-white rounded-2xl p-5 shadow">Multi-Scale Discriminator</div> |
| <div class="bg-white rounded-2xl p-5 shadow">High-Res Output</div> |
| </div> |
| </section> |
|
|
| |
| <section> |
| <h2 class="text-3xl font-bold mb-8">Real Results from the Notebook</h2> |
| <p class="text-slate-600 mb-6 leading-7"> |
| Actual visual outputs extracted directly from the Jupyter notebook, including low-resolution inputs, |
| generated super-resolved images, and side-by-side comparisons. |
| </p> |
| <div class="grid md:grid-cols-2 gap-6"> |
| <img src="notebook_image_1.png" class="rounded-2xl shadow-lg w-full" alt="Notebook output 1"> |
| <img src="notebook_image_2.png" class="rounded-2xl shadow-lg w-full" alt="Notebook output 2"> |
| </div> |
| </section> |
| |
|
|
| <section class="grid md:grid-cols-2 gap-6"> |
| <div class="bg-white rounded-3xl p-6 shadow-lg"> |
| <h3 class="text-2xl font-bold mb-4">Generator Architecture</h3> |
| <p class="text-slate-600 leading-7"> |
| Residual blocks extract high-level features, followed by progressive upsampling layers that reconstruct |
| high-frequency details such as leaf veins, lesions, and texture patterns. |
| </p> |
| </div> |
| <div class="bg-white rounded-3xl p-6 shadow-lg"> |
| <h3 class="text-2xl font-bold mb-4">Discriminator Architecture</h3> |
| <p class="text-slate-600 leading-7"> |
| A multi-scale discriminator evaluates local and global realism, encouraging outputs that are both |
| perceptually convincing and statistically similar to real high-resolution images. |
| </p> |
| </div> |
| </section> |
|
|
| <section class="space-y-8"> |
| <h2 class="text-3xl font-bold">Key Code Snippets</h2> |
|
|
| <div class="code-card rounded-3xl p-6 shadow-xl"> |
| <h3 class="text-xl font-bold mb-3 text-white">Generator Initialization</h3> |
| <pre><code>generator = Generator(upscale_factor=4) |
| discriminator = MultiScaleDiscriminator()</code></pre> |
| </div> |
|
|
| <div class="code-card rounded-3xl p-6 shadow-xl"> |
| <h3 class="text-xl font-bold mb-3 text-white">Perceptual Loss</h3> |
| <pre><code>vgg = VGGFeatureExtractor() |
| perceptual_loss = F.l1_loss( |
| vgg(sr_images), |
| vgg(hr_images) |
| )</code></pre> |
| </div> |
|
|
| <div class="code-card rounded-3xl p-6 shadow-xl"> |
| <h3 class="text-xl font-bold mb-3 text-white">Hybrid Generator Loss</h3> |
| <pre><code>g_loss = ( |
| adv_loss + |
| lambda_l1 * l1_loss + |
| lambda_perc * perceptual_loss |
| )</code></pre> |
| </div> |
|
|
| <div class="code-card rounded-3xl p-6 shadow-xl"> |
| <h3 class="text-xl font-bold mb-3 text-white">Training Loop</h3> |
| <pre><code>for lr_imgs, hr_imgs in train_loader: |
| sr_imgs = generator(lr_imgs) |
| d_loss = train_discriminator(...) |
| g_loss = train_generator(...)</code></pre> |
| </div> |
| </section> |
|
|
| <section class="grid lg:grid-cols-2 gap-10 items-start"> |
| <div> |
| <h2 class="text-3xl font-bold mb-4">Expected Performance Trend</h2> |
| <canvas id="perfChart"></canvas> |
| </div> |
| <div class="bg-white rounded-3xl p-8 shadow-lg"> |
| <h3 class="text-2xl font-bold mb-4">Optimization Strategies</h3> |
| <ul class="space-y-3 text-slate-600"> |
| <li>• Test-Time Augmentation (TTA)</li> |
| <li>• Progressive Training</li> |
| <li>• Ensemble Methods</li> |
| <li>• Fine-Tuning Strategy</li> |
| <li>• Data Augmentation</li> |
| </ul> |
| </div> |
| </section> |
|
|
| <section> |
| <h2 class="text-3xl font-bold mb-8">Technologies Used</h2> |
| <div class="flex flex-wrap gap-3"> |
| <span class="px-4 py-2 bg-white rounded-full shadow">Python</span> |
| <span class="px-4 py-2 bg-white rounded-full shadow">PyTorch</span> |
| <span class="px-4 py-2 bg-white rounded-full shadow">Torchvision</span> |
| <span class="px-4 py-2 bg-white rounded-full shadow">PIL</span> |
| <span class="px-4 py-2 bg-white rounded-full shadow">NumPy</span> |
| <span class="px-4 py-2 bg-white rounded-full shadow">Matplotlib</span> |
| </div> |
| </section> |
| </main> |
|
|
| <footer class="border-t border-slate-200 py-10 mt-12"> |
| <div class="max-w-6xl mx-auto px-6 text-slate-500"> |
| Built from the original Jupyter notebook and optimized for deployment as a static Hugging Face Space. |
| </div> |
| </footer> |
|
|
| <script> |
| new Chart(document.getElementById('perfChart'), { |
| type: 'line', |
| data: { |
| labels: ['Epoch 10', 'Epoch 20', 'Epoch 30', 'Epoch 40', 'Epoch 50'], |
| datasets: [{ data: [0.032, 0.026, 0.021, 0.018, 0.015], tension: 0.35 }] |
| }, |
| options: { |
| plugins: { |
| legend: { display: false }, |
| tooltip: { |
| callbacks: { |
| label: function(context) { |
| return 'Validation MAE: ' + context.parsed.y.toFixed(3); |
| } |
| } |
| } |
| }, |
| scales: { |
| y: { |
| reverse: true, |
| title: { display: true, text: 'Lower is Better (MAE)' } |
| } |
| } |
| } |
| }); |
| </script> |
| </body> |
| </html> |
|
|
