Skip to content

Latest commit

 

History

History
7 lines (4 loc) · 777 Bytes

File metadata and controls

7 lines (4 loc) · 777 Bytes

Handwritten Image Recognition

The aim of this assignment is to use Optical Character Recognition (OCR) technology to predict labels from handwritten text images.

OCR is a technology that converts any kind of images containing written or printed text into a machine-readable format. Deep learning techniques were used for the handwriting recognition in a real case scenario. Some of these techniques are the Feedforward Neural Networks (FNN), the Convolutional Neural Networks (CNN) – with and without data augmentation methods – and the Convolutional Recurrent Neural Networks (CRNN).

Finally, pre-trained models (VGG-16, VGG-19, ResNet50) were implemented in order to see how a model that was already used in such topics like OCR, works in handwritten text images.