/USPTO-Object-Detection

Tensorflow - Patent Classification

Primary LanguageJupyter Notebook

CNN for Multi-Figure Identification

This project looks at how Convolutional Neural Networks (CNNs) can be used to effectively identify a series of digits from patent images that are obtained from the United States Patent Office (USPTO) Database.

Problem Statement

I am attempting to predict numbers, letters, and drawings provided by an image of patent drawings from the USPTO Database.

Libraries

I used Tensorflow and Python to build this model. Tensorboard is also utilized.

Report

A more detailed report has been written as well.

Resources and Academic Journals