## Dataset Information
This dataset contains images of front contact metallization of silicon solar cells in .PNG format.
There are 5 different classes:
1. Silicon
2. Silver
3. Voids
4. Interfacial Voids
5. Glass
Currently, the dataset consists of 77 annotated images.
#### **Instructions for Generating Predictions and Evaluating Model Performance on the Benchmark Dataset**
- Images and corresponding ground truth masks are [here][1].
- For each model, we provide an inference script to generate predictions.
- Refer to the table at the bottom for links to the inference scripts for each model.
- You need to change the paths for images and model in the script accordingly.
- Once you've generated the predictions, you can use the '[get_evaluation_metrics.py][2]' script (applicable to all models in this dataset) to assess the model's performance on the generated predictions.
- You need to change the paths for images in the script accordingly.
**Metadata**:
- **Imaging Tecnique:** Scanning Electron Microscopy (SEM)
- **Samples:**
- Cores from **monocrystalline silicon PERC** modules
Exposure Time: 4-years
Exposure Location: Cocoa, FL (FSEC test site)
- Cores from **multicrystalline silicon Al-BSF** modules
Exposure Time: 10-years (FSEC test site)
Exposure Location: Cocoa, FL
- Cores from **multicrystalline silicon Al-BSF** modules
Exposure Time: 14-years
Expossure Location: Fort Meyers, FL (Florida Gulf Coast University)
- **Exposure Type:** Field operation
- **Image Size:** 1024x768, 2048x1536
**Want to contribute? Fill up this [Google Form][3]**.
<h2>
Benchmarks
</h2>
| Model | Description | Average IoU | Library | Input Image Size | Performance Metrics | Paper | Model Inference Script | Compatible PyTorch Version |
|:------------------------------------------------:|:------:|:----------:|:--------------------:|:----------------:|:------------------------------------------:|:------------------------------------------:|:------------------------------------------:|:------------------------------------------:|
[unet.pth][4]| U-Net with 4-block encoder and decoder | 84.43%| PyTorch | 1024 x 768| Overall Accuracy, Precision, Recall, IoU <br> Class-wise Accuracy, Precision, Recall, IoU * | - | [Click Here][5] | PyTorch 2.1.2 |
[1]: https://osf.io/mwg6n/files/osfstorage
[2]: https://osf.io/download/gmdy8/
[3]: https://forms.gle/4pGRhkdw4wGNy1dY9
[4]: https://osf.io/c3vsw
[5]: https://osf.io/download/b92mj/
