Files in this item

FilesDescriptionFormat

application/pdf

application/pdfECE499-Sp2019-zhang-Yichi.pdf (3MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Handwriting recognition and robotic application
Author(s):Zhang, Yichi
Contributor(s):Levinson, Stephen E.
Subject(s):Artificial Intelligence
Machine Learning
Recognition
Abstract:Handwriting is one of the most significant communication and recording information tools in our daily life. Considering the ubiquitous role that handwriting plays in our life, machine recognition of handwriting has practical importance. The purpose of this research is to let the robot view the handwritten digits, recognize them and do tasks such as calculating simple formulas depending on the digits it sees. For the recognition part, we use the MNIST data base and apply three different algorithms from machine learning including support vector machine (SVM), K-nearest neighbors (k-NN) and deep neural network (DNN), to perform the prediction of input handwritten digits. Different features are extracted from those different methods and the testing accuracy varies. By comparing the training time and the test accuracy, we can choose the best training architecture for this problem. For the robotic part, we use the camera from the robot in the robot simulator. The robot model can visualize the images of digits and apply the pre-trained weights from previous steps to get the predictions for the digits and move its joints around to perform the given tasks such as simple mathematical calculation.
Issue Date:2019-05
Genre:Other
Type:Text
Language:English
URI:http://hdl.handle.net/2142/104054
Date Available in IDEALS:2019-06-19


This item appears in the following Collection(s)

Item Statistics