|
|
--- |
|
|
license: apache-2.0 |
|
|
datasets: |
|
|
- prithivMLmods/Face-Age-10K |
|
|
language: |
|
|
- en |
|
|
base_model: |
|
|
- google/siglip2-base-patch16-512 |
|
|
pipeline_tag: image-classification |
|
|
library_name: transformers |
|
|
tags: |
|
|
- age-detection |
|
|
- SigLIP2 |
|
|
- biology |
|
|
--- |
|
|
|
|
|
 |
|
|
|
|
|
# facial-age-detection |
|
|
|
|
|
> facial-age-detection is a vision-language encoder model fine-tuned from `google/siglip2-base-patch16-512` for **multi-class image classification**. It is trained to detect and classify human faces into **age groups** ranging from early childhood to elderly adults. The model uses the `SiglipForImageClassification` architecture. |
|
|
|
|
|
> \[!note] |
|
|
> SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features |
|
|
> [https://arxiv.org/pdf/2502.14786](https://arxiv.org/pdf/2502.14786) |
|
|
|
|
|
```py |
|
|
Classification Report: |
|
|
precision recall f1-score support |
|
|
|
|
|
age 01-10 0.9614 0.9669 0.9641 2474 |
|
|
age 11-20 0.8418 0.8467 0.8442 1181 |
|
|
age 21-30 0.8118 0.8326 0.8220 1523 |
|
|
age 31-40 0.6937 0.6683 0.6808 1010 |
|
|
age 41-55 0.7106 0.7528 0.7311 1181 |
|
|
age 56-65 0.6878 0.6646 0.6760 799 |
|
|
age 66-80 0.7949 0.7596 0.7768 653 |
|
|
age 80 + 0.9349 0.8343 0.8817 344 |
|
|
|
|
|
accuracy 0.8225 9165 |
|
|
macro avg 0.8046 0.7907 0.7971 9165 |
|
|
weighted avg 0.8226 0.8225 0.8223 9165 |
|
|
``` |
|
|
|
|
|
 |
|
|
|
|
|
--- |
|
|
|
|
|
## Label Space: 8 Classes |
|
|
|
|
|
``` |
|
|
Class 0: age 01-10 |
|
|
Class 1: age 11-20 |
|
|
Class 2: age 21-30 |
|
|
Class 3: age 31-40 |
|
|
Class 4: age 41-55 |
|
|
Class 5: age 56-65 |
|
|
Class 6: age 66-80 |
|
|
Class 7: age 80 + |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## Install Dependencies |
|
|
|
|
|
```bash |
|
|
pip install -q transformers torch pillow gradio hf_xet |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## Inference Code |
|
|
|
|
|
```python |
|
|
import gradio as gr |
|
|
from transformers import AutoImageProcessor, SiglipForImageClassification |
|
|
from PIL import Image |
|
|
import torch |
|
|
|
|
|
# Load model and processor |
|
|
model_name = "prithivMLmods/facial-age-detection" # Update with actual model name on Hugging Face |
|
|
model = SiglipForImageClassification.from_pretrained(model_name) |
|
|
processor = AutoImageProcessor.from_pretrained(model_name) |
|
|
|
|
|
# Updated label mapping |
|
|
id2label = { |
|
|
"0": "age 01-10", |
|
|
"1": "age 11-20", |
|
|
"2": "age 21-30", |
|
|
"3": "age 31-40", |
|
|
"4": "age 41-55", |
|
|
"5": "age 56-65", |
|
|
"6": "age 66-80", |
|
|
"7": "age 80 +" |
|
|
} |
|
|
|
|
|
def classify_image(image): |
|
|
image = Image.fromarray(image).convert("RGB") |
|
|
inputs = processor(images=image, return_tensors="pt") |
|
|
|
|
|
with torch.no_grad(): |
|
|
outputs = model(**inputs) |
|
|
logits = outputs.logits |
|
|
probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist() |
|
|
|
|
|
prediction = { |
|
|
id2label[str(i)]: round(probs[i], 3) for i in range(len(probs)) |
|
|
} |
|
|
|
|
|
return prediction |
|
|
|
|
|
# Gradio Interface |
|
|
iface = gr.Interface( |
|
|
fn=classify_image, |
|
|
inputs=gr.Image(type="numpy"), |
|
|
outputs=gr.Label(num_top_classes=8, label="Age Group Classification"), |
|
|
title="Facial Age Detection", |
|
|
description="Upload a face image to estimate the age group: 01–10, 11–20, 21–30, 31–40, 41–55, 56–65, 66–80, or 80+." |
|
|
) |
|
|
|
|
|
if __name__ == "__main__": |
|
|
iface.launch() |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## Intended Use |
|
|
|
|
|
`facial-age-detection` is designed for: |
|
|
|
|
|
* **Demographic Analytics** – Estimate age distributions in image datasets for research and commercial analysis. |
|
|
* **Access Control & Verification** – Enforce age-based access in digital or physical environments. |
|
|
* **Retail & Marketing** – Understand customer demographics in retail spaces through camera-based analytics. |
|
|
* **Surveillance & Security** – Enhance people classification systems by integrating age detection. |
|
|
* **Human-Computer Interaction** – Adapt experiences and interfaces based on user age. |
