defectdetection / README.md
nazlicanto's picture
creation of the space
4e707ae
|
raw
history blame
1.54 kB
metadata
title: Defectdetection
emoji: 🚀
colorFrom: blue
colorTo: yellow
sdk: streamlit
sdk_version: 1.27.2
app_file: app06.py
pinned: false
license: mit

🛠️ PCB Defect Detection App

This app allows users to upload PCB images and detect defects using state-of-the-art machine learning models.

🌟 Features

  • Image Upload: Easily upload your PCB images and get instant defect predictions.
  • Visualization: Visualize the detected defects on the PCB image.
  • Defect Types: The app can identify multiple types of defects and highlight them uniquely for easy identification.

🚀 Usage

1️⃣ Uploading an Image:

  • Click on the "Browse files" button.
  • Select a PCB image from your device.
  • Sit back and relax! Let the model churn through the image and present its findings.

2️⃣ Interpreting Results:

  • It will display the original image alongside the predicted defect mask.
  • Different defect types will be highlighted using unique grayscale values.

Model Details

The app utilzes the Segformer model trained on a custom PCB dataset. The model has been fine-tuned to detect:

  • Incorrect Installation
  • Short Circuit
  • Dry Joints

... commonly found defects in PCBs.

📜 Requirements

The app is built using Streamlit and leverages the Hugging Face Transformers library for model inference. For a full list of requirements, refer to the requirements.txt file.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference