The team investigated whether eye tracking could help understand the behavior of respondents during surveys. | Photo credit: FILE PHOTO
The International Institute of Information Technology Bangalore (IIIT-B) is developing a low-cost eye-tracking tool that uses a regular webcam to detect when people answering long surveys are distracted or mentally fatigued, a step toward improving the reliability of data from public health and social surveys.
The project is funded by Machine Intelligence and Robotics CoE (MINRO).
Large-scale health and behavioral surveys are mostly conducted through door-to-door visits, with interviewers using questionnaires to interview respondents. These surveys are important for shaping public policy, but they are expensive and difficult to conduct. During the Covid-19 pandemic, such in-person surveys have become difficult, prompting agencies to rely more on online surveys. However, researchers have found that self-administered surveys often have high rates of dropouts and incomplete responses, especially when respondents feel overwhelmed.
A project led by IIIT-B professor Jaya Sreevalsan Nair with Beryl Gnanaraj, a PhD candidate at the institute, addresses this gap. The team investigated whether eye tracking could help understand the behavior of respondents during surveys. The goal is to identify signs of cognitive overload, such as loss of focus or prolonged hesitation, which often lead to questions being skipped or unreliable answers. By marking such moments, survey organizers can better assess data quality or redesign surveys to reduce respondent fatigue, said Prof. Nair.
In addition to health research, the researchers believe the technology could be adapted for other uses, including assessing reading ability in children with learning disabilities, tracking attention on digital learning platforms, detecting malpractice during online examinations, and exploring applications in mental health research.
Currently, eye tracking solutions rely on specialized hardware that is largely developed in the United States or Europe and is extremely expensive. According to the researchers, a professional eye tracking device along with its supporting software can cost close to ₹50,000. This cost barrier was a key reason for designing a solution that works with standard web cameras.
The system uses webcam footage to generate visual maps showing where the respondent is likely to be looking on the screen. These visualizations help identify points of view, which are then analyzed using computer models to estimate levels of mental effort and attention. The goal is to understand whether the respondent is concentrating, distracted, or having trouble with a particular question without the need for any additional equipment.
To do this, researchers use advanced machine learning techniques, including deep learning models, to estimate gaze direction from webcam images. Unlike professional eye trackers that directly capture precise eye movements, web camera-based systems have to work with noisier data. To improve reliability, the IIIT-B model uses both raw webcam footage and processed visual cues, a method the team says differs from most existing webcam-based trackers.
The system is currently designed for post-survey analysis, meaning that the webcam footage is reviewed after the survey is completed. Although real-time analysis is not yet part of the project, it may be explored in future phases. Currently, the team has performed a qualitative evaluation using visual outputs and found the results to be encouraging, although a formal comparison of accuracy with professional eye-tracking devices has yet to be done.
The tool has not yet been tested in a live survey setting. The team is in talks with the National Institute of Mental Health and Neurosciences (NIMHANS), Bengaluru, to first use the system in a research environment before testing it in real-world survey conditions.
One of the main challenges in the development of the tool was the creation of a large training data set. Annotating video data frame-by-frame is time-consuming, and creating enough examples to reflect real-world conditions such as poor lighting, head movements, different facial angles, and low-resolution web cameras required considerable effort.
The researchers also emphasized the importance of privacy, given that the system involves recording facial video and eye gaze information. The project is being assessed by the institute’s review board and the data is currently used only for research purposes. Any future sharing or disclosure of data will comply with applicable legal and ethical guidelines.
Published – 03 Jan 2026 20:44 IST
