GazeParser

GazeParser is a free, open-source computer tool used to study where people look. It provides a low-cost way for researchers to track eye movements and analyze visual attention without buying expensive commercial equipment.

Created by researcher Hiroyuki Sogo, the system is designed to be highly customizable and runs on multiple computer platforms. Key Components of GazeParser

GazeParser handles the entire eye-tracking process by splitting the work into two main parts:

Data Recording: It uses a standard video camera to capture video of a person’s eyes. It records data with less than 1 millisecond of error in timing, making it highly accurate for fast eye movements.

Data Analysis: Once the eye movements are recorded, GazeParser processes the data to find specific visual behaviors. How it Analyzes Visual Attention

When humans look around, our eyes do not move smoothly. GazeParser is built to look for two main types of eye movements to understand what captures a person’s attention:

Fixations: These are the moments when the eyes stay still on one spot. A fixation shows that a person is focusing on and processing a specific piece of information.

Saccades: These are the quick, jerky jumps the eyes make when moving from one focus point to another. GazeParser measures how far these jumps are and how long they take. Why Researchers Use GazeParser

Works with Python: The entire library is written in Python. This allows it to work seamlessly with popular psychology experiment software like ⁠PsychoPy and VisionEgg.

Low Cost: Commercial eye-trackers can cost thousands of dollars. GazeParser allows university labs and independent researchers to run experiments using simple webcams or affordable cameras.

Proven Accuracy: In performance tests, GazeParser performed just as well as high-end commercial systems when tracking complex vision tasks. It can pinpoint exactly where a person is looking within 0.7 to 1.2 degrees of accuracy. If you are planning an eye-tracking project, tell me:

Are you trying to set up an experiment or just analyze existing data? Do you already have a specific camera you want to use?

I can provide the right Python code steps to help you get started. Springer Nature Link