Notes Separator

Notes Separator

Inspiration

The inspiration for this project comes from the need to efficiently and accurately separate handwritten notes from other spam messages that are received on communication platforms such as WhatsApp. At the end of the semester, we students receive many messages containing various types of content, including notes, memes, jokes, and other irrelevant messages. It can be a tedious and time-consuming task to manually sort through all these messages and find the important notes.

What it does

This project aims to classify images as containing handwritten notes or not (spam) using a pre-trained convolutional neural network (CNN) model. The CNN model is trained on a large dataset of images with and without handwritten notes and is used to predict the class of new input images. The purpose of this project is to automate the task of image classification for applications such as separating handwritten notes from other spam images we receive on WhatsApp

How we built it

We used TensorFlow and Keras libraries to build and train a convolutional neural network model to classify handwritten notes and non-note images. The model was trained on a large dataset of handwritten notes and non-notes images. The training dataset was augmented to prevent overfitting. After training the model, we used it to predict the class of the images in the input directory.

Challenges we ran into

One of the challenges we faced was getting the model to generalize well on a wide range of handwritten notes and non-note images. We also had to optimize the memory consumption to avoid OOM errors when running the model on large images.

Accomplishments that we're proud of

We are proud of building a model that can accurately separate handwritten notes from non-note images with a high degree of accuracy. Also we learnt a lot while working on this project.

What we learned

We learned a lot about deep learning and computer vision techniques, such as convolutional neural networks and image pre-processing. We also learned how to optimize memory consumption when running models on large images.

What's next for Notes Separator

In the future, We plan to improve the model's accuracy and extend it to be able to detect different types of notes, such as math equations and diagrams. We also plan to build a user-friendly interface to make it easier for users to use and customize the system. The source code for the model training can be used to train model to classify different classes of things. It can be converted into an Android app to make it more accessible.