2017 Fiscal Year Research-status Report
Generating artificial eye movements on a document for an objective readability measurement
Project/Area Number |
17K00276
|
Research Institution | Osaka Prefecture University |
Principal Investigator |
デンゲル アンドレアス 大阪府立大学, 研究推進機構, 客員教授 (00773574)
|
Co-Investigator(Kenkyū-buntansha) |
石丸 翔也 大阪府立大学, 研究推進機構, 客員研究員 (10788730)
黄瀬 浩一 大阪府立大学, 工学(系)研究科(研究院), 教授 (80224939)
|
Project Period (FY) |
2017-04-01 – 2020-03-31
|
Keywords | eye tracking / readability / legibility / electrodermal activity |
Outline of Annual Research Achievements |
In this research, we aim to evaluate the readability of a document on the basis of eye movements. Since it is not realistic to ask participants to read documents with eye tracking devices every time, we develop a system which synthesizes artificial eye movements on the document and utilizes them for the readability measurement. This report summarizes following progress addressed in the research plan for FY 2017.
(1) Recording eye movements on documents with different layouts: We explored the relationship between document layout and successful comprehension. By considering a scenario of reading a textbook and solving related exercises, we asked half of participants to read a document displaying text first, and the remaining number of participants to read a document displaying exercises first. From the experimental result, it has been revealed that participants in the exercises first group read the text with longer fixation duration and shorter saccade length because they might read carefully at least relevant segments to solve tasks.
(2) Assessing cognitive workload while reading documents on different media: Cognitive workload is one of the most important factors to quantify the readability of a document. We asked participants to read documents on different media (on printed and an electronic device) with wearing an eye tracker and an electrodermal activity (EDA) wristband to measure their cognitive workload. From the experimental result, we found that the media type has a significant influence on cognitive workload and understandability of the content.
|
Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The main part of the research plan for FY 2017 is to collect reading behaviors for the gaze synthesis. As listed above, we have conducted several eye tracking experiments, and this has been achieved. In addition, there is a point that has progressed beyond our expectation and a point that is behind the proposed plan. By considering both aspects together, we may state that we reached our goals. The two aspects are:
(1) Progress beyond our expectation: An EDA wristband was not included in the proposed plan. But we utilized it and found that it is useful to quantify cognitive states of a reader and the readability of a document in our future work.
(2) Progress behind the proposed plan: The output of our current gaze synthesis is not accurate enough. We need to continue increasing the number of training samples for deep learning and to consider what types of features from documents can be utilized as inputs of the network.
|
Strategy for Future Research Activity |
As synthesizing eye movements is a challenging task, we divide the task into two steps. We will start from fixation estimation because it is valid for the measurement of reading speed and attention. Then we will involve saccades to cover regressions (re-reading on a document) and skimming. Finally, we combine approaches for fixation/saccade estimation to synthesize realistic eye movements. As soon as the initial eye movements generating system and the readability measurement system are implemented, we develop demo applications, evaluate the performance, and improve the system.
|
Causes of Carryover |
Initially, we planned to order two types of eye trackers (a remote eye tracker which can be attached to a display and mobile eye tracking glasses). But when testing them at the beginning of the project, we decided to use a remote eye tracker mainly because we need accurate gaze coordinates on a document. We will use the budget to involve other sensing modalities such as an EDA wristband in the next year.
|
Research Products
(2 results)