File size: 1,535 Bytes
b0820c1
 
 
 
 
 
 
4e707ae
b0820c1
 
 
 
4e707ae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b0820c1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
---
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