Natural error correction techniques for sketch recognition
Name
862074582-MIT.pdf
Description
Full printable version
Size
9.3 MB
Format
Adobe PDF
Checksum (MD5)
4a8fb262159bb281ce10d56cecea0dcb
Author(s)
Chang, Danica H. (Danica Hill)
Advisor(s)
Randall Davis.
Date Issued
2013
Publisher
Massachusetts Institute of Technology
Abstract
Over the past few years, a plethora of tablet devices has made it very easy for users to input information by sketching as if on paper. In addition, sketch recognition systems help users convert these sketches into information that the computer understands. While lots of work has been done in developing better sketch recognizers, very little work has previously been done on how to edit the sketch once it's been drawn, whether the error is the user's or the sketch recognizer's. In response, we developed and studied intuitive methods of interacting with a sketch recognition system to correct errors made by both the recognizer and the user. The editor allows users to click and lasso to select parts of the sketch, label the selected strokes, erase by scribbling over strokes, and even overwrite errors. Letting users provide feedback to the sketch recognizer helps improve the accuracy of the sketch as well as allows the sketch recognizer's performance to improve over time.
Description
Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 55-56).
Subjects
Electrical Engineering and Computer Science.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
M.I.T. theses are protected by
copyright. They may be viewed from this source for any purpose, but
reproduction or distribution in any format is prohibited without written
permission. See provided URL for inquiries about permission.
copyright. They may be viewed from this source for any purpose, but
reproduction or distribution in any format is prohibited without written
permission. See provided URL for inquiries about permission.
Persistent DSpace Link