The wordometer - Estimating the number of words read using document image retrieval and mobile eye tracking

Kai Steven Kunze, Hitoshi Kawaichi, Kazuyo Yoshimura, Koichi Kise

Research output: Contribution to journalArticle

40 Citations (Scopus)

Abstract

We introduce the Wordometer, a novel method to estimate the number of words a user reads using a mobile eye tracker and document image retrieval. We present a reading detection algorithm which works with over 91 % accuracy over 10 test subjects using 10-fold cross validation. We implement two algorithms to estimate the read words using a line break detector. A simple version gives an average error rate of 13,5 % for 9 users over 10 documents. A more sophisticated word count algorithm based on support vector regression with an RBF kernel reaches an average error rate from only 8.2 % (6.5 % if one test subject with abnormal behavior is excluded). The achieved error rates are comparable to pedometers that count our steps in our daily life. Thus, we believe the Wordometer can be used as a step counter for the information we read to make our knowledge life healthier.

Original languageEnglish
Article number6628579
Pages (from-to)25-29
Number of pages5
JournalUnknown Journal
DOIs
Publication statusPublished - 2013
Externally publishedYes

Fingerprint

Document Retrieval
Eye Tracking
Image retrieval
Image Retrieval
test subject
Error Rate
Count
Support Vector Regression
Cross-validation
Estimate
Reading
Fold
Detector
kernel
Detectors
regression
Line
Life

Keywords

  • document image retrival
  • eye gaze
  • eyetracking
  • word count
  • wordometer

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

The wordometer - Estimating the number of words read using document image retrieval and mobile eye tracking. / Kunze, Kai Steven; Kawaichi, Hitoshi; Yoshimura, Kazuyo; Kise, Koichi.

In: Unknown Journal, 2013, p. 25-29.

Research output: Contribution to journalArticle

@article{d0257a71f9d048f887c66b6962354193,
title = "The wordometer - Estimating the number of words read using document image retrieval and mobile eye tracking",
abstract = "We introduce the Wordometer, a novel method to estimate the number of words a user reads using a mobile eye tracker and document image retrieval. We present a reading detection algorithm which works with over 91 {\%} accuracy over 10 test subjects using 10-fold cross validation. We implement two algorithms to estimate the read words using a line break detector. A simple version gives an average error rate of 13,5 {\%} for 9 users over 10 documents. A more sophisticated word count algorithm based on support vector regression with an RBF kernel reaches an average error rate from only 8.2 {\%} (6.5 {\%} if one test subject with abnormal behavior is excluded). The achieved error rates are comparable to pedometers that count our steps in our daily life. Thus, we believe the Wordometer can be used as a step counter for the information we read to make our knowledge life healthier.",
keywords = "document image retrival, eye gaze, eyetracking, word count, wordometer",
author = "Kunze, {Kai Steven} and Hitoshi Kawaichi and Kazuyo Yoshimura and Koichi Kise",
year = "2013",
doi = "10.1109/ICDAR.2013.14",
language = "English",
pages = "25--29",
journal = "Mathematical Social Sciences",
issn = "0165-4896",
publisher = "Elsevier",

}

TY - JOUR

T1 - The wordometer - Estimating the number of words read using document image retrieval and mobile eye tracking

AU - Kunze, Kai Steven

AU - Kawaichi, Hitoshi

AU - Yoshimura, Kazuyo

AU - Kise, Koichi

PY - 2013

Y1 - 2013

N2 - We introduce the Wordometer, a novel method to estimate the number of words a user reads using a mobile eye tracker and document image retrieval. We present a reading detection algorithm which works with over 91 % accuracy over 10 test subjects using 10-fold cross validation. We implement two algorithms to estimate the read words using a line break detector. A simple version gives an average error rate of 13,5 % for 9 users over 10 documents. A more sophisticated word count algorithm based on support vector regression with an RBF kernel reaches an average error rate from only 8.2 % (6.5 % if one test subject with abnormal behavior is excluded). The achieved error rates are comparable to pedometers that count our steps in our daily life. Thus, we believe the Wordometer can be used as a step counter for the information we read to make our knowledge life healthier.

AB - We introduce the Wordometer, a novel method to estimate the number of words a user reads using a mobile eye tracker and document image retrieval. We present a reading detection algorithm which works with over 91 % accuracy over 10 test subjects using 10-fold cross validation. We implement two algorithms to estimate the read words using a line break detector. A simple version gives an average error rate of 13,5 % for 9 users over 10 documents. A more sophisticated word count algorithm based on support vector regression with an RBF kernel reaches an average error rate from only 8.2 % (6.5 % if one test subject with abnormal behavior is excluded). The achieved error rates are comparable to pedometers that count our steps in our daily life. Thus, we believe the Wordometer can be used as a step counter for the information we read to make our knowledge life healthier.

KW - document image retrival

KW - eye gaze

KW - eyetracking

KW - word count

KW - wordometer

UR - http://www.scopus.com/inward/record.url?scp=84889599875&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889599875&partnerID=8YFLogxK

U2 - 10.1109/ICDAR.2013.14

DO - 10.1109/ICDAR.2013.14

M3 - Article

AN - SCOPUS:84889599875

SP - 25

EP - 29

JO - Mathematical Social Sciences

JF - Mathematical Social Sciences

SN - 0165-4896

M1 - 6628579

ER -