SRSoheil Rayatdoost

Soheil Rayatdoost

*This is an algorithmically generated author record based on publicly available information.

Soheil Rayatdoost has a total of 7 co-authors. Here are the top co-authors by number of shared publications:

  • H. D. Taghirad

    Advanced Robotics and Automated Systems (ARAS), Industrial Control Center of Excellence (ICEE), Faculty of Electrical and Computer Engineering, K. N. Toosi University of Technology, Tehran, Iran

    2 shared publications
  • Mohammad Soleymani

    University of Southern California,USC Institute for Creative Technologies

    2 shared publications
  • David Rudrauf

    University of Geneva,Department of Psychology

    1 shared publication
  • Ehsan Arbabi

    School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran

    1 shared publication
  • M. Parsapour

    Advanced Robotics and Automated Systems (ARAS), Industrial Control Center of Excellence (ICEE), Faculty of Electrical and Computer Engineering, K. N. Toosi University of Technology, Tehran, Iran

    1 shared publication
  • M. Salehian

    Advanced Robotics and Automated Systems (ARAS), Faculty of Electrical and Computer Engineering, K. N. Toosi University of Technology

    1 shared publication

Latest Publications

  • Expression-Guided EEG Representation Learning for Emotion Recognition

    by Soheil Rayatdoost, David Rudrauf, Mohammad Soleymani

    Learning a joint and coordinated representation between different modalities can improve multimodal emotion recognition. In this paper, we propose a deep representation learning approach for emotion recognition from electroencephalogram (EEG) signals guided by facial electromyogram (EMG) and electrooculogram (EOG) signals. We recorded EEG, EMG and EOG signals from 60 participants who watched 40 short videos and self-reported their emotions. A cross-modal encoder that jointly learns the features extracted from facial and ocular expressions and EEG responses was designed and evaluated on our recorded data and MAHOB-HCI, a publicly available database. We demonstrate that the proposed representation is able to improve emotion recognition performance. We also show that the learned representation can be transferred to a different database without EMG and EOG and achieve superior performance. Methods that fuse behavioral and neural responses can be deployed in wearable emotion recognition solutions, practical in situations in which computer vision expression recognition is not feasible.


  • CROSS-CORPUS EEG-BASED EMOTION RECOGNITION

    by Soheil Rayatdoost, Mohammad Soleymani

    Lack of generalization is a common problem in automatic emotion recognition. The present study aims to explore the suitability of the existing EEG features for emotion recognition and investigate the performance of emotion recognition methods across different corpora. We introduce a novel dataset which includes spontaneous emotions and was analyzed in addition to the existing datasets for cross-corpus evaluation. We demonstrate that the performance of the existing methods significantly decreases when evaluated across different corpora. The best results are obtained by a convolutional neural network fed by spectral topography maps from different bands. We provide some evidence that stimuli-related sensory information is learned by machine learning models for emotion recognition using EEG signals.


  • Effect of different partitioning strategies of face imprint on thermal face recognition

    by Vahid Alizadeh, Soheil RayatDoost, Ehsan Arbabi

    Face recognition using thermal images has received high attention by researchers in recent years. Since skin temperature can be visualized using thermographic cameras, a unique imprint can be achieved for thermal image of a person's face. In this study, we have investigated the effect of using different partitioning strategies of face imprint on thermal face recognition. First, we have proposed a simple and reliable approach for face extraction using ellipse mask. Then, the face imprint has been attained using anisotropic diffusion filter and morphological processing. The face imprint is related to distribution of blood vessels under the face skin. In the next step, the face imprint has been partitioned using different strategies including rectangular, circular and polar approaches. The number of minutiae point in each partition has been computed. Finally, a SVM classifier based on one-against-one approach has been used for face recognition. This paper presents a framework to find the best partitioning approach resulting in highest performance. Based on the results, the rectangular partitioning was found as the best partitioning approach for face imprint, due to its highest performance (92.27%).


  • Position based sliding mode control for visual servoing system

    by M. Parsapour, S. RayatDoost, H. D. Taghirad

    This paper presents a nonlinear controller for visual servoing system. Pose estimation is one of the fundamental issues in position-based visual servoing (PBVS) approach. A few researches have focused on controller synthesis under modeling uncertainty and measurement noise of estimated position. In this research, PD-type sliding surface is designed for tracking target. The control signal is obtained from the sliding surface and the stability of the algorithm is verified by Lyapunov theory. Moreover, a recent designed robust estimator based on unscented Kalman observer (UKO) cascading with Kalman filter (KF) is used to estimate the pose, velocity and acceleration of the target. The combination of the implemented estimator and the proposed controller provide a stable and robust structure in PBVS. The reported experimental results, verify the effectiveness of the proposed method in an industrial visual servoing system.


  • Robust unscented Kalman filter for visual servoing system

    by M. Salehian, S. RayatDoost, H. D. Taghirad

    This paper presents a robust pose estimator for visual servoing system. Although various filters has been used as pose estimators, very limited research has been focused on the stability and robustness of pose estimators. UKF or EKF based pose estimator is one of most celebrated approaches in uncertain and noisy environment for nonlinear observations. However convergence of these filters is subject to some restrictive conditions in practice. In order to obtain a robust converging filter, pose estimation problem in visual servoing system is decomposed to an unscented Kalman observer (UKO) in cascade with a Kalman filter (KF). This structure inverts an uncertain nonlinear estimation problem to a certain nonlinear estimation in addition to an uncertain linear estimation. Additionally, a modified principal component analysis (PCA) based feature extractor is extended in this paper, which is shown to be robust in a noisy environment. The reported experimental results verify the effectiveness of the proposed structure in visual servoing system.