Application of a Strong Tracking Finite-Difference Extended Kalman Filter to Eye Tracking

Zhang, Jiashu; Zhang, Zutao

doi:10.1007/11816157_146

Cited by 3 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Zhang et al proposed a nonlinear unscented Kalman filter for gaze estimation [7], which can overcome the difficulties of nonlinear gaze estimation and improve the accuracy of gaze estimation. Jiashu Zhang et al proposed to use several groups of points to match the posterior probability density function of eye movement [8], which is more accurate than the estimation effect of the traditional Kalman filter. Although nonlinear Kalman filtering algorithms can improve the accuracy of gaze estimation, they are numerically unstable in practical applications and require more computational time.…”

Section: Related Workmentioning

confidence: 99%

Gaze Estimation Method Combining Facial Feature Extractor with Pyramid Squeeze Attention Mechanism

Wei

et al. 2023

Electronics

View full text Add to dashboard Cite

To address the issue of reduced gaze estimation accuracy caused by individual differences in different environments, this study proposes a novel gaze estimation algorithm based on attention mechanisms. Firstly, by constructing a facial feature extractor (FFE), the method obtains facial feature information about the eyes and locates the feature areas of the left and right eyes. Then, the L2CSNet (l2 loss + cross-entropy loss + softmax layer network), which integrates the PSA (pyramid squeeze attention), is designed to increase the correlation weights related to gaze estimation in the feature areas, suppress other irrelevant weights, and extract more fine-grained feature information to obtain gaze direction features. Finally, by integrating L2CSNet with FFE and PSA, FPSA_L2CSNet was proposed, which is fully tested on four representative publicly available datasets and a real-world dataset comprising individuals of different backgrounds, lighting conditions, nationalities, skin tones, ages, genders, and partial occlusions. The experimental results indicate that the accuracy of the gaze estimation model proposed in this paper has been improved by 13.88%, 11.43%, and 7.34%, compared with L2CSNet, FSE_L2CSNet, and FCBA_L2CSNet, respectively. This model not only improves the robustness of gaze estimation but also provides more accurate estimation results than the original model.

show abstract

Section: Related Workmentioning

confidence: 99%

Gaze Estimation Method Combining Facial Feature Extractor with Pyramid Squeeze Attention Mechanism

Wei

et al. 2023

Electronics

View full text Add to dashboard Cite

show abstract

“…[42], [43], and [44]) clearly reveal the significant features of the classical approach as follows: 1) Due to the use of the notions as "the state of the system" and "the state propagation equation" this approach is strongly based on analytical system modeling; 2) It emphasizes the need of (possibly) complete information on, and control of the actual state of the system (the notions as "observability", and "controllability" as well as the general efforts for constructing Kalman Filters or Extended Kalman Filters that are widely used model-based state estimators (e.g. [45], [46]) evidently highlight this fact); 3) Though being "liberated from the confines of the frequency domain" till it is strictly related to linear, linearizable, or fully or partly linearized state space models that can be caught in the consistent use of the terminology as "Linear Time-Invariant (LTI) Systems", "Linear Parameter Varying (LPV) Systems", and "Quasi Linear Parameter Varying (qLPV) Systems" in the control technology of our days.…”

Section: Implications For Control Technologymentioning

confidence: 99%

Cognitive Control initiative

Tar

Rudas

Kosi

et al. 2012

2012 IEEE 3rd International Conference on Cognitive Infocommunications (CogInfoCom)

View full text Add to dashboard Cite

In this paper, we make an initial effort to define the scope and goals of "Cognitive Control (CoCo)", which intends to be a research initiative for bringing about a novel generation of control systems which -in a certain sensereflect the behavior of humans while solving everyday tasks. Very briefly stated, through the use of a rich inventory of available modeling techniques, control theory has achieved profound results in model-based control approaches. However, these solutions do not seem to begin to approach the true level of human intelligence. We are convinced that by the synthesis of these "tools" from a particular point of view the great efficiency of human intelligence can be integrated/fused with the complementary virtues of our machines (e.g. high computing power, fast operation, etc.), and our CoCo initiative can be realized and applied in practice.

show abstract

Tackling complexity and missing information in adaptive control by fixed point transformation-based approach

Tar

Rudas

Nádai

et al. 2016

2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC)

View full text Add to dashboard Cite

Application of a Strong Tracking Finite-Difference Extended Kalman Filter to Eye Tracking

Cited by 3 publications

References 10 publications

Gaze Estimation Method Combining Facial Feature Extractor with Pyramid Squeeze Attention Mechanism

Gaze Estimation Method Combining Facial Feature Extractor with Pyramid Squeeze Attention Mechanism

Cognitive Control initiative

Tackling complexity and missing information in adaptive control by fixed point transformation-based approach

Contact Info

Product

Resources

About