Mayank022/Cropped_Face_Dataset_128x128

FaceCrop-12k Dataset

FaceCrop-12k is a curated dataset consisting of 12,000 high-quality human face images, each resized to a resolution of 128×128 pixels. The dataset is suitable for training and evaluating models in face generation, facial recognition, identity embedding extraction, and other computer vision tasks involving facial data.

---

output.zip/
├── 1.jpg
├── 2.jpg
├── ...
└── 12000.jpg

---

Usage Example

from PIL import Image
import os

folder = "path_to_extracted_zip"
for filename in sorted(os.listdir(folder)):
    if filename.endswith(".jpg"):
        img = Image.open(os.path.join(folder, filename))
        # Further processing here

---

Applications

• Training generative models (e.g., GANs, VAEs, diffusion models)
• Facial embedding learning and identity clustering
• Super-resolution, denoising, and image restoration tasks
• Research on facial similarity and reconstruction

---