A5083371732- Cell Image Analysis Techniques
- Medical Image Segmentation Techniques
- Image Processing Techniques and Applications
- 3D Printing in Biomedical Research
- Advanced MRI Techniques and Applications
- Advanced Neural Network Applications
- Wireless Communication Networks Research
- Domain Adaptation and Few-Shot Learning
- Neuroscience and Neuropharmacology Research
- Brain Tumor Detection and Classification
- Radiomics and Machine Learning in Medical Imaging
- Advanced Wireless Communication Techniques
- Advanced Fluorescence Microscopy Techniques
- AI in cancer detection
- Neural dynamics and brain function
- Video Surveillance and Tracking Methods
- Fetal and Pediatric Neurological Disorders
- QR Code Applications and Technologies
- Neonatal and fetal brain pathology
- Handwritten Text Recognition Techniques
- Cognitive Radio Networks and Spectrum Sensing
- Power Systems Fault Detection
- Smart Parking Systems Research
- Islanding Detection in Power Systems
- Single-cell and spatial transcriptomics
École Polytechnique Fédérale de Lausanne
2020-2025
Burke Medical Research Institute
2023-2025
Cornell University
2023-2025
Weill Cornell Medicine
2023-2025
Air University
2024
ETH Zurich
2019-2021
University of Zurich
2020-2021
Ahsanullah University of Science and Technology
2021
Laboratory for Social and Neural Systems Research
2019
University of the Punjab
2009-2010
It is critical to quantitatively analyse the developing human fetal brain in order fully understand neurodevelopment both normal fetuses and those with congenital disorders. To facilitate this analysis, automatic multi-tissue segmentation algorithms are needed, which turn requires open databases of segmented brains. Here we introduce a publicly available database 50 manually pathological non-pathological magnetic resonance volume reconstructions across range gestational ages (20 33 weeks)...
SUMMARY In recent years, we and others have identified a number of enhancers that, when incorporated into rAAV vectors, can restrict the transgene expression to particular neuronal populations. Yet, viral tools access manipulate fine subtypes are still limited. Here, performed systematic analysis single cell genomic data identify enhancer candidates for each cortical interneuron subtypes. We established set enhancer-AAV that highly specific distinct populations striatal cholinergic neurons....
Abstract Mapping the structure of mammalian brain at cellular resolution is a challenging task and one that requires capturing key anatomical features appropriate level analysis. Although neuroscientific methods have managed to provide significant insights micro macro level, in order obtain whole-brain analysis meso-scopic approach. A number can be currently used detect count cells, with, nevertheless, limitations when analyzing data high complexity. To overcome some these constraints, we...
Abstract Vasocative-intestinal-peptide (VIP + ) and somatostatin (SST interneurons are involved in modulating barrel cortex activity perception during active whisking. Here we identify a developmental transition point of structural functional rearrangements onto these around the start sensation at P14. Using vivo two-photon Ca 2+ imaging, find that before P14, both interneuron types respond stronger to multi-whisker stimulus, whereas after P14 their responses diverge, with VIP cells losing...
<title>Abstract</title> Skilled movement relies on descending cortical projections to the brainstem and spinal cord. While corticospinal neurons (CSN) have long been recognized for their role in fine motor control, contribution of remains poorly understood. Here, we identify a previously unrecognized direct cortico-brainstem circuit that emerges early development persists into adulthood. A subset subcerebral projection (SCPN) limit from earliest stages axon extension without ever extending...
Understanding the complex three-dimensional structure of cells is crucial across many disciplines in biology and especially neuroscience. Here, we introduce a set models including 3D transformer (SwinUNetR) novel self-supervised learning method (WNet3D) designed to address inherent complexity generating ground truth data quantifying nuclei volumes. We developed Python package called CellSeg3D that provides access these Jupyter Notebooks napari GUI plugin. Recognizing scarcity high-quality...
With the increasing number of vehicles on road, it has become difficult to properly control flow traffic especially in large cities like Lahore. The proposed system uses a digital camera mounted stepper motor sense road. decision open lane is done after monitoring load. heavily loaded side turned for longer time. Thus intelligent because not using fixed time frame. also integrated with GSM, thus signals can be controlled help mobile case an emergency. MATLAB <sup...
Finding a code to unravel the population of neural responses that leads distinct animal behavior has been long-standing question in field neuroscience. With recent advances machine learning, it is shown hierarchically Deep Neural Networks (DNNs) perform optimally decoding unique features out complex datasets. In this study, we utilize power DNN explore computational principles mammalian brain by exploiting Neuropixel data from Allen Brain Institute. We decode mouse visual cortex predict...
Abstract Understanding the complex three-dimensional structure of cells is crucial across many disciplines in biology and especially neuroscience. Here, we introduce a novel 3D self-supervised learning method designed to address inherent complexity quantifying volumes, often cleared neural tissue. We offer new mesoSPIM dataset show that CellSeg3D can match state-of-the-art supervised methods. Our contributions are made accessible through Python package with full GUI integration napari.
Abstract We introduce an AI-driven approach for robust 3D brain image registration, addressing challenges posed by diverse hardware scanners and imaging sites. Our model trained using SSIM-driven loss function, prioritizes structural coherence over voxel-wise intensity matching, making it uniquely to inter-scanner intra-modality variations. This innovative end-to-end framework combines global alignment non-rigid registration modules, specifically designed handle structural, intensity, domain...
Abstract With increasing neuroimaging modalities and data diversity, mapping brain regions to a standard atlas template has become challenging problem. Machine learning in general deep learning, particular, have been providing robust solutions for several tasks, including image registration segmentation. However, these methods require large amount of groundtruth labels, annotated by human experts, which is time-consuming. In this work, we introduce NeuroAtlas, an AI-based framework...
In electrical power systems, ensuring a reliable, precise, and efficient relay strategy is crucial for safe trustworthy operation, especially in multi-loop distribution systems. Overcurrent relays (OCRs) have emerged as effective solutions these challenges. This study focuses on optimizing the coordination of OCRs to minimize overall operational time main relays, thereby reducing outages. The optimization problem addressed by adjusting multiplier setting (TMS) using War Strategy Optimization...
A cheap guidance system based on AT89C52 microcontroller is developed for facilitating visually impaired people. Ultrasonic sensors are used to calculate distance of the obstacles around blind person guide user towards available path. Output in form voice which can hear e.g., right, left etc. In its advance mode, will be able recognize objects using image processing algorithms. Results presented show validity and performance system.
This paper describes the implementation of a neural network based hurdle avoidance controller for car like robot using low cost single chip 89C52 microcontroller. The is multilayer feed-forward with back propagation training algorithm. trained offline tangent-sigmoid as activation function neurons and implemented in real time piecewise linear approximation function. Results have shown that up-to twenty hidden layer can be deployed proposed technique vehicle tested various environments...
Technological advances in medical imaging are enabling us to understand healthcare datasets great detail. Machine Learning enabled methods, specifically, deep neural networks continuously achieving benchmark performances terms of accuracy and computational efficiency. However, the lack agreed-upon standard procedures, variations devices by different vendors, artifacts induced physical phenomenon sensors make data inconsistent noisy. These detrimental performance learning-based methods. In...
An automated means for predicting the virus is of utmost importance to help medical personnel detect patients, prepare reports and produce results fast impeccably so that people can get early treatment prevent future transmissions. In this work, we proposed a COVID19 detection method using chest x-ray images by training testing pre-trained deep neural network models, such as VGG19, InceptionV3, Densenet201 individually, got an accuracy 96.9%, 95.2%, 96.7% respectively. Then bolster...
This paper proposes and implements an integrated library management system for book search placement tasks. Book searching task is carried out with the help of a mobile phone. The user can check availability particular by title or author name using his information received contains code number location inside shelf. involves placing shelf according to assigned image processing. Digital camera mounted at base captures serial tag pasted on book. After has been processed, automatically placed...
Image registration is a widely-used technique in analysing large scale datasets that are captured through various imaging modalities and techniques biomedical such as MRI, X-Rays, etc. These typically collected from sites under different protocols using variety of scanners. Such heterogeneity the data collection process causes inhomogeneity or variation intensity (brightness) noise distribution. variations play detrimental role performance image registration, segmentation detection...
Abstract Two inhibitory cell types involved in modulating barrel cortex activity and perception during active whisking adult mice, are the VIP + SST interneurons. Here we identify a developmental transition point of structural functional rearrangements onto these interneuron around start sensation at P14. Using vivo two-photon Ca 2+ imaging, find that before P14, both respond stronger to multi-whisker stimulus, whereas after P14 their responses diverge, with cells losing preference neurons...
Understanding the complex three-dimensional structure of cells is crucial across many disciplines in biology and especially neuroscience. Here, we introduce a novel 3D self-supervised learning method designed to address inherent complexity quantifying volumes, often cleared neural tissue. We offer new mesoSPIM dataset show that CellSeg3D can match state-of-the-art supervised methods. Our contributions are made accessible through Python package with full GUI integration napari.
Understanding the complex three-dimensional structure of cells is crucial across many disciplines in biology and especially neuroscience. Here, we introduce a novel 3D self-supervised learning method designed to address inherent complexity quantifying volumes, often cleared neural tissue. We offer new mesoSPIM dataset show that CellSeg3D can match state-of-the-art supervised methods. Our contributions are made accessible through Python package with full GUI integration napari.