YJYaqub Jonmohamadi

Yaqub Jonmohamadi

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Yaqub Jonmohamadi has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Supervised Scene Illumination Control in Stereo Arthroscopes for Robot Assisted Minimally Invasive Surgery

    by Shahnewaz Ali, Yaqub Jonmohamadi, Yu Takeda, Jonathan Roberts, Ross Crawford, Ajay K. Pandey

    Minimally invasive surgery (MIS) offers many advantages to patients but it also imposes limitations on surgeons ability, as no tactile or haptic feedback is available. From medical robotics perspective, visualizations issues specific to MIS such as limited field of view and the lack of automatic exposure control of the surgical area make it challenging when it comes to tracking tissue, tools and camera pose as well as in perceiving depth. Lighting plays an important role in 3D reconstruction and variations due to internal illumination conditions are known to degrade vital visual information. In this work, we describe a supervised adaptive light control system to solve some of the image visualization problems of MIS. Our proposed method is able to classify underexposed and over-exposed frames and adjust lighting condition automatically to enrich image quality. Our method uses support vector machines to classify different illumination conditions. Visual feedback is provided by gradient information to assess image quality and justify classifier decision. The output of this system has been tested against two cadaver knee experiment data with an overall accuracy of 97.75% for under-exposed and 89.11% for over-exposed classes. Hardware implementation of this classifier is expected to result in adaptive lighting for robot assisted surgery as well as in providing support to surgeons by freeing them from manual adjustments to lighting controls.


  • Self-supervised Depth Estimation to Regularise Semantic Segmentation in Knee Arthroscopy

    by Fengbei Liu, Yaqub Jonmohamadi, Gabriel Maicas, Ajay K. Pandey, Gustavo Carneiro


  • Extraction of Common Task Features in EEG-fMRI Data Using Coupled Tensor-Tensor Decomposition

    by Yaqub Jonmohamadi, Suresh Muthukumaraswamy, Joseph Chen, Jonathan Roberts, Ross Crawford, Ajay Pandey

    The fusion of simultaneously recorded EEG and fMRI data is of great value to neuroscience research due to the complementary properties of the individual modalities. Traditionally, techniques such as PCA and ICA, which rely on strong non-physiological assumptions such as orthogonality and statistical independence, have been used for this purpose. Recently, tensor decomposition techniques such as parallel factor analysis have gained more popularity in neuroimaging applications as they are able to inherently contain the multidimensionality of neuroimaging data and achieve uniqueness in decomposition without making strong assumptions. Previously, the coupled matrix–tensor decomposition (CMTD) has been applied for the fusion of the EEG and fMRI. Only recently the coupled tensor–tensor decomposition (CTTD) has been proposed. Here for the first time, we propose the use of CTTD of a 4th order EEG tensor (space, time, frequency, and participant) and 3rd order fMRI tensor (space, time, participant), coupled partially in time and participant domains, for the extraction of the task related features in both modalities. We used both the sensor-level and source-level EEG for the coupling. The phase shifted paradigm signals were incorporated as the temporal initializers of the CTTD to extract the task related features. The validation of the approach is demonstrated on simultaneous EEG-fMRI recordings from six participants performing an N-Back memory task. The EEG and fMRI tensors were coupled in 9 components out of which seven components had a high correlation (more than 0.85) with the task. The result of the fusion recapitulates the well-known attention network as being positively, and the default mode network working negatively time-locked to the memory task.


  • Open Access

    Automatic Segmentation of Multiple Structures in Knee Arthroscopy Using Deep Learning

    by Yaqub Jonmohamadi, Yu Takeda, Fengbei Liu, Fumio Sasazawa, Gabriel Maicas, Ross Crawford, Jonathan Roberts, Ajay K. Pandey, Gustavo Carneiro

    Minimally invasive surgery (MIS) is among the preferred procedures for treating a number of ailments as patients benefit from fast recovery and reduced blood loss. The trade-off is that surgeons lose direct visual contact with the surgical site and have limited intra-operative imaging techniques for real-time feedback. Computer vision methods as well as segmentation and tracking of the tissues and tools in the video frames, are increasingly being adopted to MIS to alleviate such limitations. So far, most of the advances in MIS have been focused on laparoscopic applications, with scarce literature on knee arthroscopy. Here for the first time, we propose a new method for the automatic segmentation of multiple tissue structures for knee arthroscopy. The training data of 3868 images were collected from 4 cadaver experiments, 5 knees, and manually contoured by two clinicians into four classes: Femur, Anterior Cruciate Ligament (ACL), Tibia, and Meniscus. Our approach adapts the U-net and the U-net++ architectures for this segmentation task. Using the cross-validation experiment, the mean Dice similarity coefficients for Femur, Tibia, ACL, and Meniscus are 0.78, 0.50, 0.41, 0.43 using the U-net and 0.79, 0.50, 0.51, 0.48 using the U-net++. While the reported segmentation method is of great applicability in terms of contextual awareness for the surgical team, it can also be used for medical robotic applications such as SLAM and depth mapping.


  • Open Access

    Extraction of common task features in EEG-fMRI data using coupled tensor-tensor decomposition

    by Yaqub Jonmohamadi, Suresh Muthukumaraswamy, Joseph Chen, Jonathan Roberts, Ross Crawford, Ajay Pandey

    The fusion of simultaneously recorded EEG and fMRI data is of great value to neuroscience research due to the complementary properties of the individual modalities. Traditionally, techniques such as PCA and ICA, which rely on strong strong non-physiological assumptions such as orthogonality and statistical independence, have been used for this purpose. Recently, tensor decomposition techniques such as parallel factor analysis have gained more popularity in neuroimaging applications as they are able to inherently contain the multidimensionality of neuroimaging data and achieve uniqueness in decomposition without imposing strong assumptions. Previously, the coupled matrix-tensor decomposition (CMTD) has been applied for the fusion of the EEG and fMRI. Only recently the coupled tensor-tensor decomposition (CTTD) has been proposed. Here for the first time, we propose the use of CTTD of a 4th order EEG tensor (space, time, frequency, and participant) and 3rd order fMRI tensor (space, time, participant), coupled partially in time and participant domains, for the extraction of the task related features in both modalities. We used both the sensor-level and source-level EEG for the coupling. The phase shifted paradigm signals were incorporated as the temporal initializers of the CTTD to extract the task related features. The validation of the approach is demonstrated on simultaneous EEG-fMRI recordings from six participants performing an N-Back memory task. The EEG and fMRI tensors were coupled in 9 components out of which 7 components had a high correlation (more than 0.85) with the task. The result of the fusion recapitulates the well-known attention network as being positively, and the default mode network working negatively time-locked to the memory task.

  • Constrained temporal parallel decomposition for EEG-fMRI fusion

    by Yaqub Jonmohamadi, Anna Forsyth, Rebecca McMillan, Suresh D Muthukumaraswamy

    Multimodal neuroimaging has become a common practice in neuroscience research. Simultaneous EEG-fMRI is a popular multimodal recording approach due to the complementary spatiotemporal relationship between the two modalities. Several data fusion techniques have been proposed in the literature for EEG-fMRI fusion, including joint-ICA and parallel-ICA frameworks. Previous EEG-fMRI fusion approaches have used sensor-level EEG features. Recently, we introduced source-space ICA for EEG-MEG source reconstruction and component identification, which was shown to be a superior alternative to sensor-space ICA. Here, we extend source-space ICA to the fusion of EEG-fMRI data. Additionally, we incorporate the use of a paradigm signal and a lag-based signal decomposition approach to accommodate recent findings demonstrating the potentially variable lag structure between electrophysiological and BOLD signals. We evaluated this method on simulated concurrent EEG-fMRI during a boxcar task design, as well as real concurrent EEG-fMRI data from three participants performing an N-Back working memory task. The block diagram of the algorithm and corresponding source codes are provided. Based on the results of the real working memory task, for all three subjects, one frontal theta component, and one right posterior alpha component had the highest contribution coefficients (~0.5) to the paradigm-related fused component. There were also two more alpha band components with contribution coefficients of 0.3. The highest contributing fMRI component (~0.8) was one known in the literature to be related to the attention network. The second fMRI component was related to the well-known default mode network, with a contribution coefficient of 0.3.


  • Multi-band component analysis for EEG artifact removal and source reconstruction with application to gamma-band activity

    by Yaqub Jonmohamadi, Suresh D Muthukumaraswamy

    Independent component analysis (ICA) has been used extensively for artifact removal and reconstruction of neuronal time-courses in electroencephalography (EEG). Typically, ICA is applied on wide-band EEG (for example 1 to 100 Hz or similar ranges). EEG captures the activities of a large number of sources and the as the number of the components separated by the ICA is limited by the number of the sensors, only the stronger sources (in terms of magnitude and duration) will be detected by the ICA, and the activity of weaker sources will be lost or scattered amongst the stronger components. Because of the 1/f nature of the EEG spectra this biases components to the lower frequency ranges. Here we used multi-band ICA; a versatile combination of a filter bank, PCA, and ICA, to increase the number of the ICA components substantially and improve the SNR of the separated components. Using band-pass filtering we break the original signal mixture into several subbands, then using PCA we reduce the dimensionality of each subband, before applying ICA to a matrix containing all the principal components from each band. Using simulated sources and real EEG from participants, we demonstrate that multi-band ICA is able to outperform the traditional wide-band ICA in terms of both signal-to-noise ratio of the separated sources and the number of the identified independent components. We successfully separated the gamma-band neuronal components that were time-locked to a visual stimulus, as well as weak sources which are not detectable by wide-band ICA.


  • Open Access

    Multi-band component analysis for EEG artifact removal and source reconstruction with application to gamma-band activity

    by Yaqub Jonmohamadi, Suresh Daniel Muthukumaraswamy

    Independent component analysis (ICA) has been used extensively for artifact removal and reconstruction of neuronal time-courses in electroencephalography (EEG). Typically, ICA is applied on wide-band EEG (for example 1 to 100 Hz or similar ranges). Since EEG captures the activities of a large number of sources and the fact that number of the components separated by the ICA is limited by the number of the sensors, only the stronger sources (in terms of magnitude and duration) will be detected by the ICA, and the activity of weaker sources will be lost or scattered amongst the stronger components. Because of the 1/f nature of the EEG spectra this biases components to the lower frequency ranges. Here we used a versatile combination of a filter bank, PCA and ICA, calling it multi-band ICA, to both increase the number of the ICA components substantially and improve the SNR of the separated components. Using band-pass filtering we break the original signal mixture into several subbands, and using PCA we reduce the dimensionality of each subband, before applying ICA to a matrix containing all the principal components from each band. Using simulated sources and real EEG of participants, we demonstrate that multi-band ICA is able to outperform the traditional wide-band ICA in terms of both signal-noise ratio of the separated sources and the number of the identified independent components. We successfully separated the gamma-band neuronal components time-locked to a visual stimulus, as well as weak sources which are not detectable by wide-band ICA.

  • Source-space ICA for MEG source imaging

    by Yaqub Jonmohamadi, Richard D Jones

    Objective. One of the most widely used approaches in electroencephalography/magnetoencephalography (MEG) source imaging is application of an inverse technique (such as dipole modelling or sLORETA) on the component extracted by independent component analysis (ICA) (sensor-space ICA + inverse technique). The advantage of this approach over an inverse technique alone is that it can identify and localize multiple concurrent sources. Among inverse techniques, the minimum-variance beamformers offer a high spatial resolution. However, in order to have both high spatial resolution of beamformer and be able to take on multiple concurrent sources, sensor-space ICA + beamformer is not an ideal combination. Approach. We propose source-space ICA for MEG as a powerful alternative approach which can provide the high spatial resolution of the beamformer and handle multiple concurrent sources. The concept of source-space ICA for MEG is to apply the beamformer first and then singular value decomposition + ICA. In this paper we have compared source-space ICA with sensor-space ICA both in simulation and real MEG. The simulations included two challenging scenarios of correlated/concurrent cluster sources. Main Results. Source-space ICA provided superior performance in spatial reconstruction of source maps, even though both techniques performed equally from a temporal perspective. Real MEG from two healthy subjects with visual stimuli were also used to compare performance of sensor-space ICA and source-space ICA. We have also proposed a new variant of minimum-variance beamformer called weight-normalized linearly-constrained minimum-variance with orthonormal lead-field. Significance. As sensor-space ICA-based source reconstruction is popular in EEG and MEG imaging, and given that source-space ICA has superior spatial performance, it is expected that source-space ICA will supersede its predecessor in many applications.