In this study, we present an unsupervised methodology for detecting moving targets using Capella Space's latest generation SAR sensor. The sensor has the capability to dwell on a target for an extended period of t...
详细信息
ISBN:
(纸本)9798350360332;9798350360325
In this study, we present an unsupervised methodology for detecting moving targets using Capella Space's latest generation SAR sensor. The sensor has the capability to dwell on a target for an extended period of time in its spotlight (SP) mode, which we take advantage of to track moving objects in an acquisition. An approach that combines a signal-processing-based workflow with an image-domain-based one is presented. From one side, the long-dwell SAR image is exploit by doing an interfeometric processing of different azimuth sub-apertures from the long-dwell. From another side, a kernelized cross correlation technique is used. By combining intermediate results from these complementary workflows, a smooth and robust track is obtained on the targets. The algorithm is demonstrated on a long-dwell spotlight obtained over a busy shipping channel with watercraft both small and large successfully tracked.
A noise-corrupted image often requires interpolation. Given a linear denoiser and a linear interpolator, when should the operations be independently executed in separate steps, and when should they be combined and joi...
详细信息
ISBN:
(纸本)9798350344868;9798350344851
A noise-corrupted image often requires interpolation. Given a linear denoiser and a linear interpolator, when should the operations be independently executed in separate steps, and when should they be combined and jointly optimized? We study joint denoising / interpolation of images from a mixed graph filtering perspective: we model denoising using an undirected graph, and interpolation using a directed graph. We first prove that, under mild conditions, a linear denoiser is a solution graph filter to a maximum a posteriori (MAP) problem using an undirected graph smoothness prior, while a linear interpolator is a solution to a MAP problem using a directed graph smoothness prior. Next, we study two variants of the joint interpolation / denoising problem: a graph-based denoiser followed by an interpolator has an optimal separable solution, while an interpolator followed by a denoiser has an optimal non-separable solution. Experiments show that our joint denoising / interpolation method outperformed separate approaches noticeably.
This study presents a novel non-invasive method for detecting and classifying neuro-degenerative diseases such as Parkinson's (PD) and Alzheimer's (AD) through automatic speech analysis and artificial intellig...
详细信息
ISBN:
(纸本)9798350351491;9798350351484
This study presents a novel non-invasive method for detecting and classifying neuro-degenerative diseases such as Parkinson's (PD) and Alzheimer's (AD) through automatic speech analysis and artificial intelligence. The analysis of the voice recordings was carried out using different parametric extraction methods based on the MFCC and prosodic coefficients (VOT, Jitter, Shimmer, HNR, ... ) followed by a classification step based on CNN and FC-DNN neural network. These methods made it possible to extract relevant speech parameters and use them for training and classification. The results obtained showed vocal disturbances in mild and preclinical stages of PD and AD such as articulation, prosody and rhythmic abilities. Developed machine learning algorithms were able to detect subjects with PD with 98% accuracy from rapid syllable repetitions and 96% accuracy for subjects with AD from voice parameters.
Implicit Neural Representations (INRs) are a novel paradigm for signal representation that have attracted considerable interest for image compression. INRs offer unprecedented advantages in signal resolution and memor...
详细信息
ISBN:
(纸本)9798350349405;9798350349399
Implicit Neural Representations (INRs) are a novel paradigm for signal representation that have attracted considerable interest for image compression. INRs offer unprecedented advantages in signal resolution and memory efficiency, enabling new possibilities for compression techniques. However, the existing limitations of INRs for image compression have not been sufficiently addressed in the literature. In this work, we explore the critical yet overlooked limiting factors of INRs, such as computational cost, unstable performance, and robustness. Through extensive experiments and empirical analysis, we provide a deeper and more nuanced understanding of implicit neural image compression methods such as Fourier Feature Networks and Siren. Our work also offers valuable insights for future research in this area.
This study explores the feasibility and impact of using AI image generation tools as a medium for industrial designers to create visual stimuli for product design. Twenty industrial design students participated, divid...
详细信息
High-dimensional imageanalysis, such as Hyperspectral Imaging (HSI) data, poses unique challenges due to their high dimensionality and non-Euclidean structures, making their analysis and classification complex. In th...
详细信息
ISBN:
(纸本)9798350351491;9798350351484
High-dimensional imageanalysis, such as Hyperspectral Imaging (HSI) data, poses unique challenges due to their high dimensionality and non-Euclidean structures, making their analysis and classification complex. In this study, we explore the use of both graph deep learning (GDL) and multi-view graph representation learning for HSI classification. Furthermore, we present our proposed approach of multi-view Graph Convolutional Networks (GCNs) and how it leverages multiple views of the data by combining spectral and spatial features to improve classification accuracy. We discuss then specific challenges encountered when training our model on large HSIs, including managing large-scale graph data. We also discuss promising opportunities to overcome these challenges. By highlighting the challenges and opportunities associated with GDL and multi-view GCN usage for HSI classification, this study aims to shed light on recent developments and future prospects in this rapidly evolving field.
Terahertz images are widely used in the field of security check because they have obvious advantages in perspective ability compared with traditional visible light images. However, terahertz images also have the defec...
详细信息
ISBN:
(纸本)9798400716171
Terahertz images are widely used in the field of security check because they have obvious advantages in perspective ability compared with traditional visible light images. However, terahertz images also have the defects of low resolution and poor image quality, which bring considerable difficulties in identifying hidden objects. Aiming at terahertz security images, this paper designs a transformer-based registration method between terahertz images and visible light images. Firstly, the YOLOv5 network is used to detect the hidden items in the terahertz image, and the attention module is embedded in the network to improve the recognition performance;then the visible image is semantically segmented using the FCN network to extract the contours of the human body in it, and finally the transformer module is used to perform the image registration work, and the location of the detected hidden items in the terahertz image is displayed on the visible image. In this paper, experimental validation is carried out using the actual collected terahertz security image dataset, and the results verify the feasibility of the method adopted in this paper.
In this paper, we experimentally demonstrate an outdoor optical camera communications (OCC) system utilizing a wearable light-emitting diode (LED) as the transmitter. We explore the practicality of employing commercia...
详细信息
ISBN:
(纸本)9798350348750;9798350348743
In this paper, we experimentally demonstrate an outdoor optical camera communications (OCC) system utilizing a wearable light-emitting diode (LED) as the transmitter. We explore the practicality of employing commercial devices, such as an LED strip and a smartphone, in OCC links for simultaneous monitoring and communication purposes. In particular, a strip of red-green-blue (RGB) LEDs is modulated to transmit data for user identification via visible light. Each color (red, green, blue and yellow) serves as an indicator of the user's status. Our system exhibits potential applications in high-risk environments where monitoring the physical well-being of individuals is crucial.
Examining video characteristics, particularly leveraging filters such as Canny and calculating image standard deviation, prior to the video coding process is a crucial pre-processing step that enhances the efficiency ...
详细信息
ISBN:
(纸本)9798350351491;9798350351484
Examining video characteristics, particularly leveraging filters such as Canny and calculating image standard deviation, prior to the video coding process is a crucial pre-processing step that enhances the efficiency and quality of the coding workflow. By applying filters like Canny, valuable insights into the spatial distribution of edges are gained, enabling precise identification of regions with varying complexity. Simultaneously, the calculation of standard deviation provides a quantitative measure of pixel intensity variations, aiding in the assessment of overall image texture. This meticulous analysis, as proposed by our method, enables the optimization of coding parameters, including the selection of suitable compression algorithms and the refinement of spatial and temporal redundancy reduction strategies. Ultimately, the strategic incorporation of filter-based analyses in the preprocessing phase not only refines video coding processes but also lays the foundation for improved compression outcomes, addressing the unique characteristics of each video sequence with precision.
Antenna scan type recognition can enhance cognition in electronic warfare, and the airborne multi-function radar (MFR) mainly use electronic scan type. This paper considers the electronic scan type recognition problem...
详细信息
ISBN:
(纸本)9798350350920
Antenna scan type recognition can enhance cognition in electronic warfare, and the airborne multi-function radar (MFR) mainly use electronic scan type. This paper considers the electronic scan type recognition problem of airborne MFR. Specifically, it commences with an investigation of the difference between electronic scan types on pulse amplitude. Then, a multi-dimensional feature recognition method based on connected component analysis is proposed to achieve better recognition accuracy. The method first extracts the primary and secondary wave bit sequences with connected component analysis and then proposes multi-dimensional feature for recognition. Simulation results show that the proposed method demonstrates superior performance compared to traditional methods.
暂无评论