A5102920751- Image and Signal Denoising Methods
- Non-Invasive Vital Sign Monitoring
- Advanced Vision and Imaging
- Medical Image Segmentation Techniques
- Optical Coherence Tomography Applications
- Adversarial Robustness in Machine Learning
- Advanced Neural Network Applications
- Remote-Sensing Image Classification
- Anomaly Detection Techniques and Applications
- Advanced Image and Video Retrieval Techniques
- Photoacoustic and Ultrasonic Imaging
- Image Processing Techniques and Applications
- Human Pose and Action Recognition
- Advanced Image Processing Techniques
- Advanced Image Fusion Techniques
- Video Surveillance and Tracking Methods
- AI in cancer detection
- Hemodynamic Monitoring and Therapy
- Domain Adaptation and Few-Shot Learning
- Heart Rate Variability and Autonomic Control
- Image Retrieval and Classification Techniques
- Robotics and Sensor-Based Localization
- Evolutionary Algorithms and Applications
- Video Analysis and Summarization
- Industrial Vision Systems and Defect Detection
University of Waterloo
2015-2024
University of Saskatchewan
2019-2024
Broadcom (United States)
2023
Artificial Intelligence in Medicine (Canada)
2019-2020
Research Institute for Aging
2016-2019
University of Guelph
2019
Continental (Germany)
2018
10X Genomics (United States)
2016-2017
Samsung (South Korea)
2016
Seoul National University
2016
Abstract Characterizing the transcriptome of individual cells is fundamental to understanding complex biological systems. We describe a droplet-based system that enables 3′ mRNA counting tens thousands single per sample. Cell encapsulation, up 8 samples at time, takes place in ∼6 min, with ∼50% cell capture efficiency. To demonstrate system’s technical performance, we collected data from ∼250k across 29 samples. validated sensitivity and its ability detect rare populations using lines...
Recent work has shown that the murine BRCA2 tumor suppressor protein interacts with RAD51 protein. This interaction suggests participates in DNA repair. Residues 3196–3232 of were to be involved this interaction. Here, we report detailed mapping additional domains are interactions between human homologs these two proteins. Through yeast two-hybrid and biochemical assays, demonstrate specifically eight evolutionarily conserved BRC motifs encoded exon 11 brca2 a similar motif found...
Early detection of lung cancer can help in a sharp decrease the mortality rate, which accounts for more than 17% percent total related deaths. A large number cases are encountered by radiologists on daily basis initial diagnosis. Computer-aided diagnosis (CAD) systems assist offering "second opinion" and making whole process faster. We propose CAD system uses deep features extracted from an auto encoder to classify nodules as either malignant or benign. use 4303 instances containing 4323...
The recent integration of attention mechanisms into segmentation networks improves their representational capabilities through a great emphasis on more informative features. However, these ignore an implicit sub-task semantic and are constrained by the grid structure convolution kernels. In this paper, we propose novel squeeze-and-attention network (SANet) architecture that leverages effective (SA) module to account for two distinctive characteristics segmentation: i) pixel-group attention,...
We discuss entanglement of multiparticle quantum systems. propose a potential measure type pure states n qubits, the n-tangle. For system two qubits n-tangle is equal to square concurrence, and for systems three it ''residual entanglement''. show that -tangle, also generalization concurrence squared even n, use this fact prove an monotone. However, undefined odd n>3. Finally we related mixed state find analytical formula n.
<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> This paper presents the Automatic Registration of Remote-Sensing Images (ARRSI); an automatic registration system built to register satellite and aerial remotely sensed images. The is designed specifically address problems associated with images obtained at different times and/or from sensors. ARRSI capable handling geometrically distorted by various transformations such as translation, rotation,...
Journal Article Sequence organization and cytological localization of the minor satellite mouse Get access A.K.C. Wong, Wong Department Medical Biochemistry Anatomy, University CalgaryCalgary, Alberta, Canada Search for other works by this author on: Oxford Academic PubMed Google Scholar J.B. Rattner * To whom correspondence should be addressed Nucleic Acids Research, Volume 16, Issue 24, 23 December 1988, Pages 11645–11661, https://doi.org/10.1093/nar/16.24.11645 Published: 1988 history...
This paper presents an attribute clustering method which is able to group genes based on their interdependence so as mine meaningful patterns from the gene expression data. It can be used for grouping, selection, and classification. The partitioning of a relational table into subgroups allows small number attributes within or across groups selected analysis. By attributes, search dimension data mining algorithm reduced. reduction especially important in because such typically consist huge...
This paper introduces a novel approach to the problem of image inpainting through use nonlocal-means. In traditional techniques, only local information around target regions are used fill in missing information, which is insufficient many cases. More recent techniques based on concept exemplar-based synthesis utilize nonlocal but very limited way. proposed algorithm, we from multiple samples within image. The contribution each sample reconstruction pixel determined using an weighted...
An important image post-processing step for optical coherence tomography (OCT) images is speckle noise reduction. Noise in OCT multiplicative nature and difficult to suppress due the fact that addition component, also carries structural information about imaged object. To address this issue, a novel reduction algorithm was developed. The projects imaging data into logarithmic space general Bayesian least squares estimate of noise-free found using conditional posterior sampling approach....
This paper presents a quantitative radiomics feature model for performing prostate cancer detection using multiparametric MRI (mpMRI). It incorporates novel tumor candidate identification algorithm to efficiently and thoroughly identify the regions of concern constructs comprehensive detect tumorous regions. In contrast conventional automated classification schemes, this radiomics-based aims ground its decisions in way that can be interpreted understood by diagnostician. is done grouping...
ABSTRACT Characterizing the transcriptome of individual cells is fundamental to understanding complex biological systems. We describe a droplet-based system that enables 3′ mRNA counting up tens thousands single per sample. Cell encapsulation in droplets takes place ∼6 minutes, with ∼50% cell capture efficiency, 8 samples at time. The speed and efficiency allow processing precious while minimizing stress cells. To demonstrate system′s technical performance its applications, we collected data...
In this paper, we address the hyperspectral image (HSI) classification task with a generative adversarial network and conditional random field (GAN-CRF)-based framework, which integrates semisupervised deep learning probabilistic graphical model, make three contributions. First, design four types of convolutional transposed layers that consider characteristics HSIs to help extracting discriminative features from limited numbers labeled HSI samples. Second, construct networks (GANs) alleviate...
Melanoma is the deadliest form of skin cancer. Incidence rates melanoma have been increasing, especially among non-Hispanic white males and females, but survival are high if detected early. Due to costs for dermatologists screen every patient, there a need an automated system assess patient's risk using images their lesions captured standard digital camera. One challenge in implementing such locating lesion image. A novel texture-based segmentation algorithm proposed. set representative...
An automatic method for segmenting skin lesions in conventional macroscopic images is presented. The are acquired with cameras, without the use of a dermoscope. Automatic segmentation from very challenging problem due to factors such as illumination variations, irregular structural and color presence hair, well occurrence multiple unhealthy regions. To address these factors, novel iterative stochastic region-merging approach employed segment regions corresponding images, where region merging...
Single image super-resolution (SISR) is a challenging ill-posed problem which aims to restore or infer high-resolution from low-resolution one. Powerful deep learning-based techniques have achieved state-of-the-art performance in SISR; however, they can underperform when handling images with non-stationary degradations, such as for the application of projector resolution enhancement. In this paper, new UNet architecture that able learn relationship between set degraded and their...
Robotic leg prostheses and exoskeletons can provide powered locomotor assistance to older adults and/or persons with physical disabilities. However, the current locomotion mode recognition systems being developed for automated high-level control decision-making rely on mechanical, inertial, neuromuscular sensors, which inherently have limited prediction horizons (i.e., analogous walking blindfolded). Inspired by human vision-locomotor system, we an environment classification system computer...
In this work, we present a cloud-based system for non-contact, real-time recognition and monitoring of physical activities walking periods within domestic environment. The proposed employs standalone Internet Things (IoT)-based millimeter wave radar devices deep learning models to enable autonomous, free-living activity gait analysis. To train models, utilize range-Doppler maps generated from dataset real-life in-home activities. performance several is evaluated based on accuracy prediction...