Image may be NSFW.
Clik here to view.
The ACM Symposium on Eye Tracking Research & Applications (ETRA) 2024 was a hybrid conference with the in-person event at the Glasgow Grosvenor Hotel, Glasgow, UK, and virtual attendees joining via Webex Events. The conference took place from June 4 to June 7, 2024. We (Yasasi and Gavindya) attended the paper sessions and keynotes of the conference virtually.
ETRA symposium focuses on all aspects of eye movement research across a wide range of disciplines and brings together computer scientists, engineers, and behavioral scientists to enhance eye-tracking research and applications. This year, Yasasi from ODU served on the organizing committee of the ETRA conference as the publicity chair.
🎉#ETRA2024 kicks off today at #Glasglow! We are looking forward to a week full of exciting presentations and insightful discussions on groundbreaking eye tracking research! 🚀
— ETRA (@ETRA_conference) June 4, 2024
Stay tuned for updates! 👀#ETRA#eyetracking#research#conferencepic.twitter.com/KaBhL0IKp6
Keynote 1
Prof. Joy Hirsch, a Professor of Psychiatry, Comparative Medicine, and Neuroscience at Yale School of Medicine, delivered the first keynote of ETRA 2024. Her keynote speech was titled "Eye Movements During Live Face-to-Face Interactions: A Social Information Processing Stage". She highlighted the impact of using functional near-infrared spectroscopy (fNIRS), an optical imaging technology, for neuroimaging during live interpersonal interactions. She discussed how fNIRS, combined with eye tracking, EEG, and other data streams, helps bridge the gap between controlled lab settings and real-world social interactions. She argued that their findings on the link between eye movements and active visual sensing may drive the neural dynamics of human social interactions.
Kicking off the Day 2 of @ETRA_conference 2024 with the keynote by Prof. Joy Hirsch from Yale University @YaleMed. In her speech, she discusses "Eye Movements During Live Face-to-Face Interactions: A social Information Processing Stage"#ETRA2024#Glasgowpic.twitter.com/vzPjp6RVk1
— Yasasi (@Yasasi_Abey) June 5, 2024
Paper Session 1: Visual Attention
The first paper session of the conference "Visual Attention" started after the first keynote and was chaired by Dr. Minoru Nakayama from the Tokyo Institute of Technology. This session included four full papers and three short papers. This session began with Dr. Oleg Komogortsev from Texas State University and Meta Reality Labs presenting their paper, "Per-Subject Oculomotor Plant Mathematical Models and the Reliability of Their Parameters".
Paper Session 1: Visual Attention @ETRA_conference 2024 just started. @olegkomo from @RealityLabs and @txst is now presenting their paper on "Per-Subject Oculomotor Plant Mathematical Models and the Reliability of Their Parameters" at #ETRA2024.
— Yasasi (@Yasasi_Abey) June 5, 2024
Paper: https://t.co/oqppbK2ITTpic.twitter.com/fZUGKnfiQK
Next, Cristina Rovira-Gay from the Technical University of Catalonia presented their short paper, "Saccade Characteristics during Fusional Vergence Tests as a Function of Vergence Demand". This study investigated the characteristics of saccades that occurred during the fusional vergence test. The study results provide valuable insights into how binocular vision and oculomotor control respond during demanding vergence situations, with potential implications for diagnosing and understanding binocular vision disorders. This paper won the Best Short Paper award at ETRA 2024.
📢 The paper titled "Saccade Characteristics during Fusional Vergence Tests as a Function of Vergence Demand" won the Best Short Paper Award at #ETRA2024 tonight! 🎉🏆
— ETRA (@ETRA_conference) June 6, 2024
Paper DOI: https://t.co/GnpPbtJOgQ@la_UPCpic.twitter.com/P8VVCsZolE
The third presenter of the visual attention session, Joseph MacInnes from Swansea University presented their full paper, "Assessing the Size of the Functional Field of View in a Gaze-Contingent Search Paradigm". This paper explores the extent of the functional field of view (FFV) when using a gaze-contingent search method. Their results suggest that an aperture larger than 12° (medium-sized aperture) does not improve search performance, and the medium-sized aperture may be optimal for the visual search task in this study. Following that, Maurice Koch from the University of Stuttgart presented the paper, "How Deep Is Your Gaze? Leveraging Distance in Image-Based Gaze Analysis". He discussed their depth-adaptive thumbnail method for varying image size according to the eye-to-object distance. This paper was awarded the Best Short Paper Honorable Mention at ETRA 2024.
"How Deep Is Your Gaze? Leveraging Distance in Image-Based Gaze Analysis" paper was awarded the Best Short Paper Honorable Mention at #ETRA2024! 🏆🎉 Congratulations, Maurice Koch and his team @Uni_Stuttgart! 👏👏 pic.twitter.com/W46vVpKUZT
— ETRA (@ETRA_conference) June 6, 2024
At the visual attention session, Yao Wang from the University of Stuttgart presented the full paper, "VisRecall++: Analyzing and Predicting Visualization Recallability from Gaze Behavior". In this paper, they proposed VisRecall++, a novel recallability dataset that contains gaze data on 200 visualizations, and GazeRecallNet, a novel computational method to predict recallability from gaze behavior that outperforms the state-of-the-art model RecallNet and three other baselines. Next, Dr. Nina A. Gehrer from the University of Tübingen and Dr. Andrew Duchowski from Clemson University presented their short paper, "Assessment of body-related attention processes via mobile eye tracking: A pilot study to validate an automated analysis pipeline". The final presentation of the visual attention session at ETRA 2024, was by Parvin Emami from the University of Luxembourg on "Impact of Design Decisions in Scanpath Modeling". This paper investigates how different design parameters affect the performance of scanpath evaluation metrics using a state-of-the-art computational model, DeepGaze++.
Paper Session 2: Cognitive Analysis and Learning
The second paper session was chaired by Heather Graz from the University of Dundee. The session included a short paper presentation and five full paper presentations. The first speaker of the session, Dr. Babette Bühler from the University of Tübingen presented their full paper, "On Task and in Sync: Examining the Relationship between Gaze Synchrony and Self-Reported Attention During Video Lecture Learning". In this study, they investigated the relationship between gaze synchrony and self-reported attention in online video-based learning and the results of this study revealed that attentive participants exhibited higher synchronization of eye movements.
Next, Anwesha Das from Saarland University presented their full paper "Shifting Focus with HCEye: Exploring the Dynamics of Visual Highlighting and Cognitive Load on User Attention and Saliency Prediction", which examines the joint impact of visual highlighting and dual-task-induced cognitive load on gaze behavior. Following that, Ziang Wu from Memorial University presented their short paper, "Measuring Motor Task Difficulty using Low/High Index of Pupillary Activity". Their work explores the effectiveness of the Low/High Index of Pupillary Activity (LHIPA) for measuring the task workloads of users in a simulated surgical task.
The next presentation at the cognitive analysis and learning session, “Automated Assessment of Eye-hand Coordination Skill using a Vertical Tracing Task on a Gaze-sensitive Human Computer Interaction Platform for Children with Autism” by Dharma Rane from the Indian Institute of Technology. This work introduces a virtual reality-based automated gaze-sensitive tool to assess eye-hand coordination skills of children with Autism quantitatively. Then, Kathrin Kennel from the University of Kaiserslautern-Landau presented their full paper, “Real-time prediction of students’ math difficulties using raw data from eye tracking and neural networks”. This work presents an algorithm that simulates a dynamic classification in a real-time adaptive learning system based on gaze data. The final presenter at paper session 2, Kuno Kurzhals from the University of Stuttgart, presented their full paper, “Teaching Eye Tracking: Challenges and Perspectives". In this study, they implemented a teaching strategy for courses on eye tracking and demonstrated that eye tracking can be effectively learned in a short time by non-experts.
@ETRA_conference 2024, Paper Session 2 started with @babette_bue from @uni_tue presenting their paper "On Task and in Sync: Examining the Relationship between Gaze Synchrony and Self-Reported Attention During Video Lecture Learning". #ETRA2024
— Gavindya (@Gavindya2) June 5, 2024
Paper: https://t.co/n9XQOrc5kMpic.twitter.com/yyMsqAmThs
Keynote 2
Prof. James M. Rehg, a Founder Professor of Computer Science and Industrial and Enterprise Systems Engineering at the University of Illinois Urbana-Champaign delivered the second keynote titled "An Egocentric Approach to Social AI". In his speech, Prof. Rehg highlighted the importance of advancing AI to address core aspects of social interaction, such as joint attention and theory of mind, through egocentric (first-person) perception. This approach has potential applications in augmented and virtual reality, as well as in improving the diagnosis and treatment of conditions like autism. He also discussed the challenges and progress in inferring social attention from multimodal sensor data and developing benchmark datasets for multimodal social understanding.
An interesting takeaway from Prof. Rehg's talk was learning about Ego4D, a collaborative project with Meta Reality Labs. In this project, they gathered egocentric perception data, integrating multimodal sensory information during social tasks, to advance fundamental AI research for multimodal machine perception in first-person video understanding. The Ego4D dataset is a massive-scale, egocentric dataset, and benchmark suite collected across 74 worldwide locations and 9 countries, with over 3,670 hours of daily-life activity video.
Day 3 of @ETRA_conference 2024 starts off with the keynote by Prof. James M. Rehg @RehgJim from @UofIllinois. In his speech, Prof. Rehg discusses the background on egocentric perception and an egocentric approach towards social #AI.#ETRA2024pic.twitter.com/RDwCLPcAHw
— Yasasi (@Yasasi_Abey) June 6, 2024
Paper Session 3: Computer Vision and Machine Learning
Session 3: Computer Vision and Machine Learning was chaired by Dr. Pawel Kasprowski from the Silesian University of Technology. First up in the paper session 3, Viet Dung Nguyen presented "Deep Domain Adaptation: A Sim2Real Neural Approach for Improving Eye-Tracking Systems". This paper is a collaboration between Rochester Institute of Technology and Meta Reality Labs. He discussed the sime2real problem that defines the mismatch between the synthetic eye images produced in a simulated environment and the real eye images. They presented a multi-step neural pipeline to address the sim2real problem in eye-tracking with domain adaptation. This paper won the Best Full Paper Award at the ETRA 2024.
🎉 Congratulations, Viet Dung Nguyen and his co-authors for wining the Best Full Paper Award at #ETRA2024! 🏆👏
— ETRA (@ETRA_conference) June 6, 2024
Paper Title: Deep Domain Adaptation: A Sim2Real Neural Approach for Improving Eye-Tracking Systems
Paper DOI: https://t.co/zoI2tbJDk0@RealityLabs@RITtigerspic.twitter.com/ZlmaLrzrgt
Next presenter, Virmarie Maquiling from the Technical University of Munich presented their full paper, "Zero-Shot Segmentation of Eye Features Using the Segment Anything Model (SAM)" at ETRA 2024 paper session 3. In this paper, they studied the ability of the Segment Anything Model to segment features from eye images captured in virtual reality setups. The third presenter of session 3, Dr. Kenan Bektas from the University of St. Gallen, presented their paper, "NeighboAR: Efficient Object Retrieval using Proximity- and Gaze-based Object Grouping with an AR System". This study developed NeighboAR, which detects objects in a user's surroundings and generates a graph that stores object proximity relationships and the user's gaze dwell times for each object. Following that, the full paper, "Using Deep Learning to Increase Eye-Tracking Robustness, Accuracy, and Precision in Virtual Reality" was presented by Kevin D. Barkevich from the Rochester Institute of Technology. This work objectively assessed how different machine learning-based methods for eye feature tracking affect gaze estimation using both feature-based and model-based approaches.
Nerea Aranjuelo from Vicomtech presented the full paper on "Learning Gaze-aware Compositional GAN from Limited Annotations". In this study, they introduced a generative framework for producing annotated gaze data by utilizing both labeled and unlabeled data sources. The final presenter of the computer vision and machine learning session, Florian Strohm from the University of Stuttgart, presented their full paper, "Learning User Embeddings from Human Gaze for Personalized Saliency Prediction". This work introduced a new approach for extracting user embeddings from pairs of natural images and their corresponding saliency maps, created using a small amount of user-specific eye-tracking data.
Paper Session 3: Computer Vision and Machine Learning @ETRA_conference 2024 just started. Viet Dung Nguyen from @RITtigers is now presenting their paper on "Deep Domain Adaptation: A Sim2Real Neural Approach for Improving Eye-Tracking Systems" at #ETRA2024. @RealityLabspic.twitter.com/Trm4dmtW0J
— Yasasi (@Yasasi_Abey) June 6, 2024
Paper Session 4: Privacy and Security
The conference has a session on privacy and security and the session was chaired by Dr. Lynsay Shepherd from Abertay University. This session started with Süleyman Özdel from the Technical University of Munich presenting their paper, "Privacy-preserving Scanpath Comparison for Pervasive Eye Tracking". To address the gap in state-of-the-art privacy-preserving eye-tracking methods that do not focus on scanpaths, this paper introduces a novel privacy-preserving scanpath comparison protocol. Next, Dr. Frode S. Volden from the Norwegian University of Science and Technology presented their paper, "My Eyes Don’t Consent! Exploring Visual Attention in Cookie Consent Interfaces" at the ETRA 2024 privacy and security session. In this work, they conducted a study to explore how users' visual attention differs across different variants of cookie consent interfaces. Following that, Valentin Foucher from Ulm University presented the full paper, “Unveiling deceptive intentions: insights from eye movements and pupil size”. This paper presents a study that contributes to a novel understanding of deceptive behaviors by distinguishing different levels of deceptive intentions; at least concealing and faking information, using eye tracking.
@ETRA_conference 2024, Paper Session 4: Privacy and Security started with Süleyman Özdel from @TU_Muenchen presenting their paper titled "Privacy-preserving Scanpath Comparison for Pervasive Eye Trackin"#ETRA2024
— Gavindya (@Gavindya2) June 6, 2024
Paper: https://t.co/MN0TMJrl5Ipic.twitter.com/0QNcSqH1K1
Paper Session 5: AR/VR and Interaction
The third day of the main conference had the final paper session on Augmented Reality/Virtual Reality (AR/VR) and Interaction. Carmen Wang from Meta chaired the session comprising seven publications. The AR/VR and Interaction session started with the presentation on "RPG: Rotation Technique in VR Locomotion using Peripheral Gaze" by Jaeyoon Lee from Korea University. In this paper, they proposed using peripheral gaze for body rotation with foveal gaze for view control, ensuring that the two tasks do not interfere with each other. Next was the full paper "GazeIntent: Adapting dwell-time selection in VR interaction with real-time intent modeling" by Anish S. Narkar from Virginia Tech. This study proposes a novel gaze-based selection architecture named GazeIntent, that uses a temporal prediction model of interaction intent to scale and adapt dwell-time thresholds for gaze-only selection.
Dr. Hans Gellersen from Lancaster University presented their full paper on "GazeSwitch: Automatic Eye-Head Mode Switching for Optimized Hands-Free Pointing" at the AR/VR and Interaction session. This paper introduces GazeSwitch, a Machine Learning-based technique that optimizes real-time switching between eye and head modes for fast and precise hands-free pointing. The following presentation was a short paper, "Eyes on the Narrative: Exploring the Impact of Visual Realism and Audio Presentation on Gaze Behavior in AR Storytelling" by Dr. Florian Weidner from Lancaster University. This study explores how auralization and visualization influence the gaze behavior of AR agents by examining how rendering style and audio type influence gaze behavior during a narrative AR experience. The full paper, "40 Years of Eye Typing: Challenges, Gaps, and Emergent Strategies" by Aleesha Hamid and Dr. Per Ola Kristensson from the University of Cambridge was presented at the AR/VR and Interaction session. This paper provides a comprehensive review and analysis of the literature, aiming to unify the diverse efforts in providing eye-typing solutions for augmentative and alternative communication users.
Dr. Alex Q. Chen from the Singapore Institute of Technology presented their paper, "Eye Strokes: An Eye-gaze Drawing System for Mandarin Characters". This is a collaborative work between the Singapore Institute of Technology,Tan Tock Seng Hospital, and the University of Glasgow. This work introduced Eye Strokes, a technique that uses eye gaze as an input modality to identify Mandarin characters for patients with motor neuron disease (MND). This method interprets eye movements as strokes in Mandarin characters to predict and identify the intended Mandarin word. The session's final presentation was the full paper on "Remapping the Document Object Model using Geometric and Hierarchical Data Structures for Efficient Eye Control" by Chris Porter and Daniel Vella from the University of Malta. This paper introduced novel interaction patterns as part of a purpose-built gaze-native web browser (Cactus), namely Quadtree-based Target Selection with Secondary Confirmation and Hierarchical Re-rendering of Navigation Menus.
Paper Session 5: AR/VR and Interaction at #ETRA2024 started with the presentation on "RPG: Rotation Technique in VR Locomotion using Peripheral Gaze" by Jaeyoon Lee from Korea University.
— Yasasi (@Yasasi_Abey) June 7, 2024
Paper DOI: https://t.co/sVKMzs4ubrpic.twitter.com/Ppx7UPfQ70
Social Events
On Wednesday, June 5th night, the conference reception was held at Glasgow City Chambers, Scotland. Glasgow City Chambers is renowned for its grand Victorian architecture, featuring one of the finest and most ornate interiors in the UK. Bailey Norman MacLeod and general chair, Dr. Mohamed Khamis gave their warm welcome to the ETRA 2024 attendees in Glasgow City Chambers.
#ETRA2024 Reception at the beautiful Glasgow City Chambers! 🏛️😍 pic.twitter.com/PmsSjjK5lx
— ETRA (@ETRA_conference) June 5, 2024
After four successful days, ETRA 2024 ended with the dinner banquet at Glasgow's City Halls and Old Fruitmarket. The dinner banquet was a delightful evening, gathering the attendees of the ETRA 2024 conference and featuring outstanding music performed by the Woodside Ceilidh Band.
#ETRA2024 Banquet took place at #Glasgow Old Fruitmarket tonight! 🍽️✨An eventful evening filled with great company from the #eyetracking community! pic.twitter.com/1ssdo5rZ4H
— ETRA (@ETRA_conference) June 6, 2024
During the banquet, general chairs, Dr. Mohamed Khamis (University of Glasgow, UK) and Dr. Yusuke Sugano (University of Tokyo, Japan), announced that the ETRA 2025 will be held in Tokyo, Japan.
See you all again at #ETRA2025 in Tokyo..! 🇯🇵 https://t.co/jvlIIzsPEh
— ETRA (@ETRA_conference) June 11, 2024