A Metric Learning Approach to Eye Movement Biometrics

Lohr, Dillon; Griffith, Henry; Aziz, Samantha; Komogortsev, Oleg V.

doi:10.1109/ijcb48548.2020.9304859

Cited by 11 publications

(10 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As mentioned in the previous related work, the feature engineering for eye tracking classification remains a main research area. In Lohr et al (2020), the authors explore using a metric learning approach to extract eye gaze features. They trained a set of three multilayer perceptrons to find fixations, saccades, and post-saccadic oscillations and reached benchmark performance for the detection.…”

Section: Related Work On Deep Learning For Eye Trackingmentioning

confidence: 99%

Imaging Time Series of Eye Tracking Data to Classify Attentional States

et al. 2021

View full text Add to dashboard Cite

It has been shown that conclusions about the human mental state can be drawn from eye gaze behavior by several previous studies. For this reason, eye tracking recordings are suitable as input data for attentional state classifiers. In current state-of-the-art studies, the extracted eye tracking feature set usually consists of descriptive statistics about specific eye movement characteristics (i.e., fixations, saccades, blinks, vergence, and pupil dilation). We suggest an Imaging Time Series approach for eye tracking data followed by classification using a convolutional neural net to improve the classification accuracy. We compared multiple algorithms that used the one-dimensional statistical summary feature set as input with two different implementations of the newly suggested method for three different data sets that target different aspects of attention. The results show that our two-dimensional image features with the convolutional neural net outperform the classical classifiers for most analyses, especially regarding generalization over participants and tasks. We conclude that current attentional state classifiers that are based on eye tracking can be optimized by adjusting the feature set while requiring less feature engineering and our future work will focus on a more detailed and suited investigation of this approach for other scenarios and data sets.

show abstract

Section: Related Work On Deep Learning For Eye Trackingmentioning

confidence: 99%

Imaging Time Series of Eye Tracking Data to Classify Attentional States

et al. 2021

View full text Add to dashboard Cite

show abstract

“…As representatives for deep learning based methods we compare against the DeepEyedentification network, which differs from DeepEyedentificationLive in that it can only process monocular data and lacks presentation-attack detection; and Abdelwahab and Landwehr [29], who train a distributional sequence embedding on raw gaze sequences and pupil dilations. Additionally, we compare against two recent representatives for deep metric learning based approaches: Whereas Lohr et al [28] learn an embedding from extracted features of scanpaths by optimizing the triplet loss, Abdelwahab and Landwehr [30] train the same distributional sequence embedding as presented in their prior study [29] end-to-end by optimizing the Wasserstein loss.…”

Section: Reference Methodsmentioning

confidence: 99%

“…In order to offer a comparison to Lohr et al and Friedman et al, we are limited to report the performance of both presented in Lohr et al [28] and to evaluate our model in the exact same evaluation protocol they use, because of unavailable source code and implementational details.…”

Section: Reference Methodsmentioning

confidence: 99%

“…Recent work uses deep learning, either processing extracted features [27], [28] or learning an embedding end-toend from the raw eye tracking signal [12], [29], [30]. Out of all prior approaches, DeepEyedentification is the only model that is able to utilize micro-movements contained in the raw signal.…”

Section: Related Workmentioning

confidence: 99%

“…Since there is no publicly available implementation of Lohr et al [28] and Friedman et al [22], we compare Deep-EyedenticationLive to published results [22], [28] in the exact same experimental setting. We perform 4-fold cross validation over the 269 subjects of the GazeBase TEX data set.…”

Section: Identification and Identity Verificationmentioning

confidence: 99%

See 2 more Smart Citations

DeepEyedentificationLive: Oculomotoric Biometric Identification and Presentation-Attack Detection Using Deep Neural Networks

Makowski

Prasse

Reich

et al. 2021

IEEE Trans. Biom. Behav. Identity Sci.

View full text Add to dashboard Cite

We study involuntary micro-movements of both eyes, in addition to saccadic macro-movements, as biometric characteristic. We develop a deep convolutional neural network that processes binocular eye-tracking signals and verifies the viewer's identity. In order to detect presentation attacks, we develop a model in which the movements are a response to a controlled stimulus. The model detects replay attacks by processing both the controlled but randomized stimulus and the ocular response to this stimulus. We acquire eye movement data from 150 participants, with 4 sessions per participant and conduct experiments on this new and legacy data sets with varying tracker precision and sampling rate. We observe that the model detects replay attacks reliably. For identification and identity verification, the model attains substantially lower error rates than prior work. We explore the relationships between training population size, training data volume, types of visual stimuli, number of training and enrollment sessions, interval between enrollment and probe sessions on one hand and the model performance on the other hand.

show abstract