Yolov8-Web-Implementation-Using-Flask

BAYBAYIN HANDWRITTEN RECOGNITION SYSTEM

Baybayin AI is a project designed to help individuals recognize handwritten baybayin characters via uploaded images.

Tools used:

- PYTHON FLASK
- Ultralytics YOLOV8 Algorithm (Detection)
- HTML & CSS
- ROBOFLOW DATASETS 

Goals

- Gather Handwritten dataset to be used in training
- Train a machine learning model to recognize characters
- Create a Web implementation of the project
- Be able to use the model to recognize images
- Predict the images correctly with above 70% Accuracy

Challenges

- Datasets used annotation and labeling
- Low accuracy prediction
- Implementation of YoloV8 Model

Dataset

All dataset used for training can be found in the link below: link: https://universe.roboflow.com/lorence-john-ejercito-w6cx7/baybayin-text-detector-9klcl/dataset/2

Algorithm

The training model used is a Convolutional neural network(CNN) called You only look once (YOLOv8) from Ultralytics. It is a computer vision algorithm used for detecting patterns and objects on images.

1. Gather Dataset, Annotate and Label the characters.
2. Divide the dataset for Training, Validation and Testing
3. Load the dataset on Google colab or Jupyter notebook
4. Train the model
5. Validate the model
6. Test the model

Collaborators

• Vashti Karmelli Camu
• Diane Mae Corcino
• Jamie Jasmine Sano
• Paul Adrian Torres (ME)