AquaSense Waste Detection Model
YOLOv8 object detection model for automated waste sorting.
Classes (5)
| ID | Name |
|---|---|
| 0 | aluminum_soda_cans |
| 1 | glass_beverage_bottles |
| 2 | paper_cups |
| 3 | plastic_soda_bottles |
| 4 | plastic_water_bottles |
Evaluation Results (client plastic bottle class)
| Metric | Value |
|---|---|
| Precision | 67.4% |
| Recall | 52.0% |
| F1-Score | 58.7% |
| AP@50 | 0.410 |
| Mean IoU (TP) | 0.832 |
Usage
from ultralytics import YOLO
from huggingface_hub import hf_hub_download
model_path = hf_hub_download(repo_id="ninditya/aquasense-waste-detection-model", filename="best.pt")
model = YOLO(model_path)
results = model.predict("your_image.jpg", conf=0.40)
Dataset
Test images and labels: ninditya/aquasense-waste-detection-dataset