« Back to Advance Program

Papers/Notes: EPIC #FAIL

Monday, April 12
11:30 AM - 1:00 PM

Estimating Residual Error Rate in Recognized Handwritten Documents Using Artificial Error Injection
Edward Lank, University of Waterloo, Canada
Ryan Stedman, University of Waterloo, Canada
Michael Terry, University of Waterloo, Canada

Describes the use of artificial errors to callibrate human performance when verifying handwriting recognition. Demonstrates that human performance on artificial errors and recognition errors is similar.

Predicting the Cost of Error Correction in Character-Based Text Entry Technologies
Ahmed S. Arif, York University, Canada
Wolfgang Stuerzlinger, York University, Canada

This article presents and verifies a new "error correction cost" model for character-based text entry technologies. It differentiates between human and system factors and enhances evaluation, comparison, and prediction.

 SHRIMP - Solving Collision and Out of Vocabulary Problems in Mobile Predictive Input with Motion Gesture
Jingtao Wang, University of California at Berkeley, U.S.A.
Shumin Zhai, IBM Almaden Research Center, U.S.A.
John Canny, University of California at Berkeley, U.S.A.

Describes an effective mobile text entry system for camera phones. It maintains the speed advantage of dictionary driven input while overcoming the collision and OOV problems without mode switching.

« Back to Advance Program