1 Introduction
2 Related work
-
Rereading, skimming, or skipping some parts of the text strongly affects the algorithms.
-
The vertical error cannot be greater than a line spacing.
-
A chin rest or a professional eye tracker is used.
3 Vertical error correction
3.1 Creating the fixation lines
3.2 Estimating and correcting coarsely the vertical error
3.3 Performing a precise (fixation line; text line) matching
4 Comparison with the state of the art
4.1 Similarity measure and global alignment
-
The difference of length between s k and t l
-
The distance between s k and t l
4.2 Differences with the proposed method
5 Experimental results
-
E1: read forward the text without rereading nor skipping any part of the text.
-
E2: read only the second paragraph of a text.
-
E3: read the two paragraphs of the text and reread the first one.
-
E4: read the second paragraph before the first one.
5.1 Ground truth
5.2 Experiment results
5.2.1 Preserved reading order: E1 and E2
E1 | E2 | |
---|---|---|
Total number of lines | 168 | 77 |
Yamaya’s method | 60 % | 83 % |
Proposed method | 69 % | 81 % |
5.2.2 Rereading and different reading order: E3 and E4
E3 | E4 | |
---|---|---|
Total number of lines | 207 | 203 |
Yamaya’s method | 64 % | 17 % |
Proposed method | 72 % | 61 % |
5.3 General discussion
Reading order | Reading order is | Total | |
---|---|---|---|
is preserved | not preserved | ||
Total number of lines | 245 | 410 | 655 |
Yamaya’s method | 67 % | 41 % | 51 % |
Proposed method | 73 % | 66 % | 69 % |