|
|
--- |
|
|
language: |
|
|
- en |
|
|
base_model: |
|
|
- Ultralytics/YOLOv8 |
|
|
pipeline_tag: object-detection |
|
|
tags: |
|
|
- planogram |
|
|
- shelf-gaps |
|
|
--- |
|
|
|
|
|
# 🛒 Retail Shelf Gap Detection Model |
|
|
|
|
|
This is a YOLOv8-based object detection model fine-tuned on the **Shelf Images for Planograms** dataset to detect **gaps** in retail shelf arrangements. The model aims to assist in **planogram compliance** checking by identifying empty spaces on shelves and computing compliance scores. |
|
|
|
|
|
## 📌 Use Case |
|
|
|
|
|
This model is intended for retail analytics teams to: |
|
|
- Detect shelf gaps in retail store images. |
|
|
- Quantify compliance against a predefined planogram. |
|
|
- Evaluate image quality and calculate gap-related metrics. |
|
|
|
|
|
## 🧠 Model Details |
|
|
|
|
|
- **Base Model:** [Ultralytics/YOLOv8](https://github.com/ultralytics/ultralytics) |
|
|
- **Task:** Object Detection |
|
|
- **Trained On:** Shelf Images for Planograms Dataset (2095 images) |
|
|
- **Framework:** PyTorch via Ultralytics |
|
|
|
|
|
## 🧪 Metrics and Scoring Breakdown |
|
|
|
|
|
The model is used as part of a retail shelf scoring pipeline that computes: |
|
|
- **Gap Score** (50%) |
|
|
- **Image Quality Score** (30%) |
|
|
- **Gap Density Score** (20%) |
|
|
- **Final Compliance Score** = Weighted sum of the above |
|
|
|
|
|
## 🛠️ How to Use |
|
|
|
|
|
```python |
|
|
from ultralytics import YOLO |
|
|
from huggingface_hub import hf_hub_download |
|
|
|
|
|
# Download model |
|
|
model_path = hf_hub_download(repo_id="akul-29/Retail-Shelf-Gap-Detection_Model", filename="best.pt") |
|
|
model = YOLO(model_path) |
|
|
|
|
|
# Run inference |
|
|
results = model("path_to_your_image.jpg") |
|
|
results.show() |
