Abstract Recently, object trackers based on Siamese networks have attracted considerable attentions due to their remarkable tracking performance and widespread application. Especially, the anchor-based methods exploit region proposal subnetwork get accurate prediction of a target make great improvement. However, those cannot capture spatial information very well pre-defined anchors will hinder robustness. To solve these problems, we propose Siamese-based anchor-free algorithm with multiscale...

Canonical polyadic decomposition (CPD) of multi-subject complex-valued fMRI data can be used to provide spatially and temporally shared components among groups with both magnitude phase information. However, the CPD model is not well formulated due large subject variability in spatial temporal modalities, as high noise level complexvalued data. Considering that shift-invariant across subjects, we propose further impose a sparsity constraint on maps denoise inter-subject well. More precisely,...

Tucker decomposition can provide an intuitive summary to understand brain function by decomposing multi-subject fMRI data into a core tensor and multiple factor matrices, was mostly used extract functional connectivity patterns across time/subjects using orthogonality constraints. However, these algorithms are unsuitable for extracting common spatial temporal subjects due distinct characteristics such as high-level noise. Motivated successful application of image denoising the intrinsic...

Abstract Spatial source phase, the phase information of spatial maps extracted from functional magnetic resonance imaging (fMRI) data by data‐driven methods such as independent component analysis (ICA), has rarely been studied. While observed shown to convey unique brain information, role in representing intrinsic activity is yet not clear. This study explores for identifying differences between patients with schizophrenia (SZs) and healthy controls (HCs) using complex‐valued resting‐state...

Convolutional neural networks (CNN) have exhibited great success in image classification. The application of CNN to classification patients with brain disorders and healthy controls is also promising using functional magnetic resonance imaging (fMRI) data. However, the shortage number subjects a challenge for training CNN. Spatial maps separated from fMRI data by independent component analysis (ICA) can provide solution this problem within an ICA-CNN framework. As such, we propose three...

Abstract Objective. Dynamic functional network connectivity (dFNC), based on data-driven group independent component (IC) analysis, is an important avenue for investigating underlying patterns of certain brain diseases such as schizophrenia. Canonical polyadic decomposition (CPD) a higher-way dynamic tensor, can offer innovative spatiotemporal framework to accurately characterize potential spatial and temporal fluctuations. Since multi-subject dFNC data from sliding-window analysis are also...

Recently, Siamese-based trackers have achieved excellent performance in object tracking. However, the high speed and deformation of objects movement process make tracking difficult. Therefore, we incorporated cascaded region-proposal-network (RPN) fusion coordinate attention into Siamese trackers. The proposed network framework consists three parts: a feature-extraction sub-network, block, RPN block.We exploit which can embed location information channel attention, to establish long-term...

Precisely segmenting the hippocampus from brain is crucial for diagnosing neurodegenerative illnesses such as Alzheimer’s disease, depression, etc. In this research, we propose an enhanced segmentation algorithm based on 3D U-Net that can significantly increase performance. First, a dynamic convolution block designed to extract information more comprehensively in steps of U-Net’s encoder and decoder. addition, improved coordinate attention applied skip connections step weight reduce...

Independent vector analysis (IVA) has exhibited great potential for the group of magnitude-only fMRI data, but rarely been applied to native complex-valued data. We propose an adaptive fixed-point IVA algorithm by taking into account extremely noisy nature, large variability source component (SCV) distribution, and non-circularity The multivariate generalized Gaussian distribution (MGGD) is exploited match SCV based on nonlinearity, shape parameter MGGD estimated using maximum likelihood...

Abstract Despite the impressive performance of correlation filter-based trackers in terms robustness and accuracy, have room for improvement. The majority existing use a single feature or fixed fusion weights, which makes it possible tracking to fail case deformation severe occlusion. In this paper, we propose multi-feature response map adaptive strategy based on consistency individual features fused feature. It is able improve accuracy by building better object appearance model. Moreover,...

Abstract The KIII model is a bionic olfactory proposed based on the physiological structure of animal system and has been applied to pattern recognition problems such as tea classification EEG recognition. To explore versatility improve its performance model, this study improved by introducing adaptive histogram equalization, Gaussian filtering, feature fusion methods gridded extraction cropping in preprocessing stage, effective image features enhances ability; Aiming at computational cost...

Tensor decomposition of fMRI data has gradually drawn attention since it can explore the multi-way data's structure which exists inherently in brain imaging. For multi-subject analysis, time shifts occur inevitably among different participants, therefore, shift-invariant tensor should be used. This method allows for arbitrary along one modality, and yield satisfactory results analyzing multi-set with datasets. In this study, we presented first application to simulated courses variations...

Magnitude-only resting-state fMRI data have been largely investigated via independent component analysis (ICA) for exacting spatial maps (SMs) and time courses. However, the native complex-valued rarely studied. Motivated by significant improvements achieved ICA of task than magnitude-only data, we present an efficient method de-noising SM estimates which makes full use data. Our two main contributions include: (1) The first application a post-ICA phase method, originally proposed to...