README.md

---
base_model:
- Ultralytics/YOLO11
pipeline_tag: object-detection
library_name: ultralytics
metrics:
- mAP50
- mAP50-95
- accuracy50
- precision
- recall
- f1
model-index:
- name: MacPaw/yolov11l-ui-elements-detection
  results:
  - task:
      type: object-detection
    metrics:
    - type: accuracy
      value: 0.65359
      name: accuracy@0.5
  - task:
      type: object-detection
    metrics:
    - type: precision
      value: 0.49055
      name: precision
  - task:
      type: object-detection
    metrics:
    - type: recall
      value: 0.43433
      name: recall
  - task:
      type: object-detection
    metrics:
    - type: f1
      value: 0.43776
      name: f1
  - task:
      type: object-detection
    metrics:
    - type: map
      value: 0.46644
      name: mAP@0.5
  - task:
      type: object-detection
    metrics:
    - type: map
      value: 0.31295
      name: mAP@0.5-0.95
datasets:
- MacPaw/Screen2AX-Element
license: agpl-3.0
---

# 🔍 YOLOv11l — UI Elements Detection

This model is a fine-tuned version of [`Ultralytics/YOLO11`](https://huggingface.co/Ultralytics/YOLO11), trained to detect **UI elements** in macOS application screenshots.

It is part of the **Screen2AX** project — a research effort focused on generating accessibility metadata using computer vision.

---

## 🧠 Task Overview

- **Task:** Object Detection
- **Target:** Individual UI elements
- **Supported Labels:**
  ```
  ['AXButton', 'AXDisclosureTriangle', 'AXImage', 'AXLink', 'AXTextArea']
  ```

This model detects common interactive components typically surfaced in accessibility trees on macOS.

---

## 🗂 Dataset

- Training data: [`MacPaw/Screen2AX-Element`](https://huggingface.co/datasets/MacPaw/Screen2AX-Element)

---

## 🚀 How to Use

### 🔧 Install Dependencies

```bash
pip install huggingface_hub ultralytics
```

### 🧪 Load the Model and Run Predictions

```python
from huggingface_hub import hf_hub_download
from ultralytics import YOLO

# Download the model
model_path = hf_hub_download(
    repo_id="MacPaw/yolov11l-ui-elements-detection",
    filename="ui-elements-detection.pt",
)

# Load and run prediction
model = YOLO(model_path)
results = model.predict("/path/to/your/image")

# Display result
results[0].show()
```

---

## 📜 License

This model is licensed under the **GNU Affero General Public License v3.0 (AGPL-3.0)**, as inherited from the original YOLOv11 base model.

---

## 🔗 Related Projects

- [Screen2AX Project](https://github.com/MacPaw/Screen2AX)
- [Screen2AX HuggingFace Collection](https://huggingface.co/collections/MacPaw/screen2ax-687dfe564d50f163020378b8)
- [YOLOv11l — UI Groups Detection](https://huggingface.co/MacPaw/yolov11l-ui-groups-detection)

---

## ✍️ Citation

If you use this model in your research, please cite the Screen2AX paper:

```bibtex
@misc{muryn2025screen2axvisionbasedapproachautomatic,
      title={Screen2AX: Vision-Based Approach for Automatic macOS Accessibility Generation}, 
      author={Viktor Muryn and Marta Sumyk and Mariya Hirna and Sofiya Garkot and Maksym Shamrai},
      year={2025},
      eprint={2507.16704},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2507.16704}, 
}
```

---

## 🌐 MacPaw Research

Learn more at [https://research.macpaw.com](https://research.macpaw.com)