https://github.com/tinh2044/readstudentcard
https://github.com/tinh2044/readstudentcard
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/tinh2044/readstudentcard
- Owner: tinh2044
- Created: 2024-07-31T03:41:55.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2024-08-10T12:30:16.000Z (10 months ago)
- Last Synced: 2025-02-07T17:20:39.266Z (4 months ago)
- Language: Jupyter Notebook
- Size: 20.7 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Read Student Card
## Tab of Content
* [Introduction](#introduction)
* [Pipeline](#pipeline)
* [DetectionObject](#detection-object)
* [OCR](#ocr)
* [Run project](#run-project)
## Introduction
- A small project applying deep learning models to retrieve important information in student cards.
- Applying YOLOv8 object detection model, build OCR image to convert text from image to text.
## Pipeline
- The student card image will be passed through the YOLOv8 model to detect the image area with text, avatar, and QR code (if any).
- Next, the image areas with text will be cut into separate images and grayscale.
- Finally, it will be passed to the OCR model to get the information.
## Detection Object
Using YOLO model to detect image area with text, avatars and QR codes
Results of the training
Testing on a sample image
## OCR
Apply CNN, Bi-LSTM ,and CTC loss architecture to build OCR model.
- Result of model
## Run Project
- Clone repo to your local machine and move to it
```Terminal
git clone [email protected]:tinh2044/ReadStudentCard.git
cd ./ReadStudentCard
```
- Create virtual environments with [conda](https://conda.io/projects/conda/en/latest/index.html) to avoid conflicts
```Terminal
conda create --name readStudentCard
conda activate readStudentCard
```
- Install requirements.
```Terminal
pip install -r ./requirements.txt
```
- Run server by command
```Terminal
uvicorn server.app.main:app --host 127.0.0.1 --port 5000
```
- Testing project with sample image, run `index.html` file in client folder