yosedie/carpet-analysis-api

🧵 Macro Carpet Classification Dataset

A high-resolution macro photography dataset consisting of 898 curated carpet images across 6 distinct carpet styles and 3 fiber material compositions, developed for automated quality assessment and authentication research.

📊 Dataset Specifications

• —Total Images: 898 high-resolution macro photographs
• —Image Dimensions: 1600 × 1600 pixels
• —Format: JPEG (.jpg)
• —Modality: Macro surface photography (01_Makroskopis)

🏷️ Annotations & Metadata

Each image is cataloged in metadata.csv with the following attributes:

• —`file_name`: Path to image file in data/
• —`style` (6 Classes):
• —Asia
• —Malaysia
• —Mie
• —Persia
• —Polos
• —Sejadah
• —`material` (3 Classes):
• —Polypropylene
• —Nylon
• —Polyester
• —`price_range`: Market price bracket (e.g., 100-200K, 200-500K)
• —`sample_id`: Sample sequence number (e.g., 001, 002)

🚀 Quickstart: Hugging Face datasets

from datasets import load_dataset

# Load dataset
dataset = load_dataset("yosedie/carpet-analysis-api")

# Inspect sample
sample = dataset["train"][0]
print("Style:", sample["style"])
print("Material:", sample["material"])
print("Image:", sample["image"])

🔗 Related Resources

• —Pre-trained Deep Learning Models: yosedie/carpet-analysis-api (Model Hub)
• —Live Inference Backend: yosedie/carpet-analysis-api (Space)
• —API Documentation: https://yosedie-carpet-analysis-api.hf.space/docs

👤 Citation & Attribution

@misc{carpet_macro_dataset,
  author = {yosedie},
  title = {Macro Carpet Classification Dataset},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/yosedie/carpet-analysis-api}}
}