CoolFace
Datasetpublic

khanusa/facial-skin-conditions

Dataset Description This dataset contains 1,103 records of SYNTHETIC facial skin analysis with real medical images, detailed Vietnamese descriptions, and conversational Q&A pairs. It's specifically designed for training multimodal AI models to analyze and discuss dermatological conditions in Vietnamese. 🖼️ Dataset Highlights 1,103 real facial skin images showing various dermatological conditions Vietnamese SYNTHETIC medical descriptions written by dermatological… See the full description on the dataset page: https://huggingface.co/datasets/khanusa/facial-skin-conditions.

sourceHugging Faceupdated 1y agoView on Hugging Face
4likes33downloads
Dataset Card

Dataset Description

This dataset contains 1,103 records of SYNTHETIC facial skin analysis with real medical images, detailed Vietnamese descriptions, and conversational Q&A pairs. It's specifically designed for training multimodal AI models to analyze and discuss dermatological conditions in Vietnamese.

🖼️ Dataset Highlights

  • —1,103 real facial skin images showing various dermatological conditions
  • —Vietnamese SYNTHETIC medical descriptions written by dermatological experts
  • —Structured medical extractions covering skin type, acne, and pigmentation
  • —Conversational Q&A pairs for training medical chatbots in Vietnamese
  • —High-quality annotations suitable for supervised learning

Dataset Structure

The dataset follows this exact structure:

json
{
  "id": "levle1_503",
  "image": "PIL.Image.Image object",
  "description": "Da mặt có biểu hiện tiết dầu nhiều, gây bóng nhờn...",
  "extractions": {
    "Tình trạng da": "Da dầu, lỗ chân lông giãn nở, bề mặt sần sùi...",
    "Tình trạng mụn": "Đa dạng bao gồm mụn đầu đen, mụn viêm...",
    "Tình trạng sắc tố": "Nhiều vết thâm đỏ và nâu (tăng sắc tố sau viêm)...",
    "Kết luận chi tiết": "Da mặt cho thấy tình trạng mụn trứng cá..."
  },
  "conversations": [
    {
      "role": "user",
      "content": "Da này loại gì?"
    },
    {
      "role": "system", 
      "content": "Da này là da dầu."
    }
  ]
}

Data Fields

FieldTypeDescription
idstringUnique identifier for each record
imagePIL.ImageActual facial skin photograph
descriptionstringDetailed Vietnamese description of skin conditions
extractionsdictStructured medical analysis with 4 key areas
extractions["Tình trạng da"]stringSkin condition assessment
extractions["Tình trạng mụn"]stringAcne/pimple condition analysis
extractions["Tình trạng sắc tố"]stringSkin pigmentation evaluation
extractions["Kết luận chi tiết"]stringDetailed medical conclusion
conversationslist[dict]Q&A pairs with roles (user/system)

Usage Example

python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("khanusa/facial-skin-conditions")

# Access a sample record
sample = dataset["train"][0]

# Get the image
image = sample["image"]  # PIL Image object
print(f"Image size: {image.size}")

# Get medical analysis
print("Skin condition:", sample["extractions"]["Tình trạng da"])
print("Acne condition:", sample["extractions"]["Tình trạng mụn"]) 
print("Pigmentation:", sample["extractions"]["Tình trạng sắc tố"])
print("Conclusion:", sample["extractions"]["Kết luận chi tiết"])

# Get conversation data
for turn in sample["conversations"]:
    print(f"{turn['role']}: {turn['content']}")

Medical Categories Covered

Skin Conditions (Tình trạng da)

  • —Da dầu (Oily skin)
  • —Da khô (Dry skin)
  • —Da hỗn hợp (Combination skin)
  • —Da nhạy cảm (Sensitive skin)
  • —Lỗ chân lông giãn nở (Enlarged pores)

Acne Types (Tình trạng mụn)

  • —Mụn đầu đen (Blackheads)
  • —Mụn đầu trắng (Whiteheads)
  • —Mụn viêm (Inflammatory acne)
  • —Mụn mủ (Pustules)
  • —Mụn ẩn (Closed comedones)
  • —Mụn nang (Cysts)

Pigmentation Issues (Tình trạng sắc tố)

  • —Tăng sắc tố sau viêm (Post-inflammatory hyperpigmentation)
  • —Tàn nhang (Freckles)
  • —Nám da (Melasma)
  • —Đốm sắc tố (Age spots)
  • —Giảm sắc tố (Hypopigmentation)