---
license: cc-by-nc-sa-4.0
extra_gated_fields:
  full_name:
    type: text
    label: What is your full name?
    required: true
  email:
    type: text
    label: What is your email address?
    required: true
  company:
    type: text
    label: Which company or institution are you affiliated with?
    required: false
  intended_use:
    type: text
    label: Please describe your intended use of this model.
    required: true
  agreement:
    type: text
    label: >-
      Type "I agree" to confirm you have read and accept the license and usage
      conditions.
    required: true
tags:
- objectdetection
- illegal_waste
- ai
---

---

![Example input](./kineva_bg_coco1.jpg)


# Model Card for `rebotnix/rb_illegal_waste`

> **Illegal Waste Detection** – Trained by KINEVA, Built by REBOTNIX, Germany
Current State: in production and re-training.

---

`rb_illegal_waste` is a specialized object detection model trained to detect **illegal waste** — improperly disposed garbage, dumped refuse, and unauthorized waste deposits found in public spaces, roadsides, forests, and urban areas. Designed for reliable detection across varying environments and lighting conditions, this model is well suited for environmental monitoring, municipal enforcement, and smart city waste management.

Developed and maintained by **REBOTNIX**, Germany, https://rebotnix.com

# About KINEVA
KINEVA is an automated training platform based on the MCP Agent system. It regularly delivers new visual computing models, all developed entirely from scratch. This approach enables the creation of customized models tailored to specific client requirements, which can be retrained and re-released as needed. The platform is particularly suited for applications that demand flexibility, adaptability, and technological precision—such as industrial image processing, smart city analytics, or automated object detection.

KINEVA is continuously evolving to meet the growing demands in the fields of artificial intelligence and machine vision. https://rebotnix.com/en/kineva

---

## Example Predictions

| Input Image | Detection Result |
|-------------|------------------|
| <img src="./example_illegal_waste_1.jpeg" width="300"/> | <a href="./rb_illegal_waste/blob/main/output_1.jpg" target="_blank"> <img src="./output_1.jpg" width="300"/></a> |
| <img src="./example_illegal_waste_2.jpeg" width="300"/> | <a href="./rb_illegal_waste/blob/main/output_2.jpg" target="_blank"> <img src="./output_2.jpg" width="300"/></a> |
| <img src="./example_illegal_waste_3.jpeg" width="300"/> | <a href="./rb_illegal_waste/blob/main/output_3.jpg" target="_blank"> <img src="./output_3.jpg" width="300"/></a> |

---

## Model Details

- **Architecture**: KINEVA SILVER *(custom training head with optimized anchor boxes)*
- **Task**: Illegal Waste Detection (single class)
- **Trained on**: Proprietary illegal waste dataset
- **Format**: PyTorch `.pth` + ONNX and TRT export available on request
- **Training Framework**: PyTorch + KINEVA + custom augmentation

---

We're happy to **license or provide access to all intermediate weights** for research or further development purposes. Please feel free to reach out.

## Dataset

The model was trained on a proprietary dataset of illegal waste images, featuring:

- Illegally dumped waste including household refuse, bulk items, and construction debris
- Varied environments including roadsides, forests, urban areas, and industrial zones
- Diverse lighting conditions including daylight, overcast, and shadows
- Multiple waste types, sizes, and levels of occlusion

---

## Intended Use

| Intended Use | Not Intended Use |
|----------------|---------------------|
| Illegal waste detection in images | Surveillance without human review |
| Environmental monitoring and compliance | Military / lethal applications |
| Municipal waste management and enforcement | Real-time safety-critical decisions without human oversight |
| Smart city cleanliness analytics | |

---

## Limitations

- May yield false positives in cluttered environments with legitimate waste containers
- Not fine-tuned for thermal or night vision imagery
- Occlusion and extreme viewing angles may reduce detection accuracy

---

## Usage Example

```python

from kineva import KINEVA

#initialize model
model = KINEVA(model="models/kineva_illegal_waste.pth")

#run inference on image
final_boxes, final_scores, final_labels = model.detect("example_illegal_waste_1.jpeg", threshold=0.35)

#draw detection
model.draw(final_boxes, final_scores, final_labels, output_path="./outputs/output_1.jpg")

```

---

## Contact

For commercial use or re-training this model support, or dataset access, contact:

**REBOTNIX**  
Email: [communicate@rebotnix.com](mailto:communicate@rebotnix.com)  
Website: [https://rebotnix.com](https://rebotnix.com)

---

## License

This model is released under **CC-BY-NC-SA** unless otherwise noted. For commercial licensing, please reach out to the contact email.

---