Please use this identifier to cite or link to this item:
Title: Development of eye gaze system for object selection in cluttered environment
Authors: Sin, Wei Zheng
Keywords: Engineering::Mechanical engineering::Mechatronics
Issue Date: 2021
Publisher: Nanyang Technological University
Source: Sin, W. Z. (2021). Development of eye gaze system for object selection in cluttered environment. Final Year Project (FYP), Nanyang Technological University, Singapore.
Project: A196
Abstract: In recent years, eye gaze or eye tracking is gaining a lot of traction, particularly in gaze-based human robot interaction (HRI). The information obtained from eye tracking can be used for a variety of purposes including but not limited to object grasping, wheelchair navigation and accessing computers or communication aids. There is a need to accurately decipher the eye gaze information collected. However, modelling eye gaze information as a fixation or fixing it onto one point is not sufficient as natural eye gaze fluctuates around an area. This issue is further exacerbated in a cluttered environment, as the user’s eye gaze is scattered across many objects. This project aims to explore a feasible solution to ensure the user is able to correctly convey their intentions for object selection in a cluttered environment while maintaining user comfort. This project uses an eye gaze system, Pupil Core from Pupil Labs and is implemented in a ROS environment. This project also utilizes a Mask R-CNN, Detectron2 for object detection. This project also makes use of the Wasserstein metric as an evaluator to compare the probability distribution between the objects’ saliency map and the user’s eye gaze distribution to predict the user’s intention for object selection.
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:MAE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
1.39 MBAdobe PDFView/Open

Page view(s)

Updated on Jan 19, 2022


Updated on Jan 19, 2022

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.