Pneumonia, an almost fatal bacterial disease generally affecting one or both lungs and brought on by the bacterium Streptococcus pneumonia, can strike any human. The World Health Organisation (WHO) reports that pneumo...
详细信息
ISBN:
(纸本)9798350391558;9798350379990
Pneumonia, an almost fatal bacterial disease generally affecting one or both lungs and brought on by the bacterium Streptococcus pneumonia, can strike any human. The World Health Organisation (WHO) reports that pneumonia accounts for one in three fatalities in India. Chest X-rays are used to detect pneumonia, require the examination of qualified radiotherapists. Establishing an automated approach to identify pneumonia would therefore be advantageous to treat the illness as soon as feasible, particularly in distant places. Convolutional Neural Networks (CNNs) have attracted a lot of interest for disease diagnosis because of the effectiveness of deep learning algorithms in evaluating medical imagery. Higher accuracy, scalability, and automation opportunities are only a few of the benefits of the suggested CNN-based approach over conventional methods. This study proposes to use deep learning methods to find intricate patterns and characteristics from medical imaging data that are suggestive of pneumonia, hence enhancing diagnosis accuracy. Additionally, picture categorization jobs benefit much from features obtained from big datasets using pre-trained CNN models. This study assesses the capability of these models in this work to diagnose abnormal and normal chest X-rays using pre-trained CNN models as feature extractors and various classifiers subsequently. There is an analytical identification of the optimal CNN model for the task.
The article discusses imageprocessingalgorithms for creating a full-fledged fire detection system based on UAV collections. The wildfire detection issue is solved using artificial intelligence: YOLOv8 model. The fir...
详细信息
ISBN:
(数字)9798331534141
ISBN:
(纸本)9798331534158
The article discusses imageprocessingalgorithms for creating a full-fledged fire detection system based on UAV collections. The wildfire detection issue is solved using artificial intelligence: YOLOv8 model. The fire classification based on the fuel map obtained using LiDAR data was also implemented. The scheme of the information system that will perform detection and classification is proposed. This information system is useful at the decision-making stage to determine the strategy and tactics of firefighting. With several improvements, the system can be used to process incoming data in real time.
In the information age, imageprocessing technology has become prevalent across various domains. To enhance image correction, computer vision algorithms can be employed. Traditional methods for structural system ident...
详细信息
ISBN:
(数字)9798350376173
ISBN:
(纸本)9798350376180
In the information age, imageprocessing technology has become prevalent across various domains. To enhance image correction, computer vision algorithms can be employed. Traditional methods for structural system identification rely on wired or wireless measurement systems for data gathering, which can be cumbersome and time-consuming due to the need for arranging sensors and data collection equipment. The proposed system involves two primary steps: line drawing extraction and rendering. During the extraction phase, an edge detection technique that integrates low and middle-level visual information is utilized. Based on the results of traditional edge detection, the edges are simplified and connected using computer vision principles. This approach treats imageprocessing as the focal point and aims to establish a scientific model for distorted images. image restoration can be facilitated through sample image filtering and noise reduction techniques. Additionally, this article delves into the principles and associated technologies of computer vision, analyzing patterns of visual interaction. Chinese traditional art imagery, as a cultural medium and symbol, fulfills an artistic communication function while preserving the nation's rich cultural heritage. The imageprocessing techniques grounded in computer vision algorithms offer numerous advantages and superior accuracy.
Steganography is the practice of hiding information by embedding it as secret data within various types of digital media to strengthen security. Numerous algorithms have been proposed for image steganography with a co...
详细信息
Constructive approaches, principles, models for optimizing the identification of micro-objects based on the use of statistical, dynamic models with mechanisms for filtering noise, foreign particles on images of medica...
详细信息
ISBN:
(数字)9798331511241
ISBN:
(纸本)9798331511258
Constructive approaches, principles, models for optimizing the identification of micro-objects based on the use of statistical, dynamic models with mechanisms for filtering noise, foreign particles on images of medical objects and pollen grains have been developed. Methods for optimizing the identification of micro-objects with the selection of characteristic objects in the image, the combination of image models with the mechanisms of segmentation, filtering, transformation, and alignment of the laser imageprocessing zone with the CAD model are proposed. algorithms for selecting the contours of objects on an image using a border detector in the transverse direction to each vector of the CAD model have been developed. Modeling of images of microsamples for medical diagnostics of chest diseases was carried out. The width, straightness of the boundaries of structural objects, local defects, parameters of regulation and synchronization of the components of the vision system are determined. algorithms for imageprocessing under conditions of a priori insufficiency, uncertainty of parameters, and low accuracy of data processing are studied. A mechanism for suppressing impulse noise and noise is implemented based on various filtering methods, preserving the boundaries of objects and small-sized parts. Mathematical expressions for estimating identification errors due to non-stationarity, inadequacy of approximation, interpolation, extrapolation of the image contourare obtained, which are synthesized with cubic, biquadratic, interpolation spline functions and wavelet transforms. A software package for recognition and classification of micro-objects has been developed.
The combination of Artificial Intelligence (AI) and Neuroscience speaks to an advantageous relationship with significant suggestions. AI, a multidisciplinary field in CS which points to imitate insights in machines, w...
详细信息
With the diversification of reconnaissance equipment and methods in the military field, optical stealth technology has become a hot spot in military research under the operational needs of optoelectronic confrontation...
详细信息
Energy efficient designs are need of the hour and imageprocessing applications are error tolerant application. Where the error present in the computing does not impact the output visual quality. This work proposes an...
详细信息
ISBN:
(数字)9798331528126
ISBN:
(纸本)9798331528133
Energy efficient designs are need of the hour and imageprocessing applications are error tolerant application. Where the error present in the computing does not impact the output visual quality. This work proposes an efficient approach to design a Bfloat floating point division. This work proposes three algorithms that are implemented at the hardware level. All the proposed algorithms will produce lesser error in the output and does not impact the output visual quality. The proposed work reduces the energy sianlficantly than the existing works.
In the imageprocessing domain, the growth of digital data has intensified the need for efficient and robust optimization techniques. This research study aims to develop and evaluate advanced optimization methods tail...
详细信息
ISBN:
(数字)9798350368413
ISBN:
(纸本)9798350368420
In the imageprocessing domain, the growth of digital data has intensified the need for efficient and robust optimization techniques. This research study aims to develop and evaluate advanced optimization methods tailored specifically for improving the performance of imageprocessing tasks. It explores the latest advancements in optimization algorithms, including evolutionary algorithms, metaheuristic approaches, and deep learning-based optimization techniques. The study provides an in-depth analysis of these methods, elucidating their strengths, weaknesses, and areas of applicability across diverse imageprocessing tasks such as image denoising, image reconstruction, image segmentation, and image enhancement. By comparing their performance through comprehensive experiments, the paper demonstrates substantial improvements in computational efficiency, accuracy, and generalization. These results highlight the potential of optimization methods to significantly enhance the quality and speed of imageprocessing pipelines, opening new avenues for breakthroughs in computer vision, medical imaging, remote sensing, and other domains. Ultimately, this research not only empowers practitioners with cutting-edge tools but also paves the way for future exploration in the application of optimization techniques within imageprocessing.
The indoor imaging visible light positioning (VLP) technology based on the image sensor (IS) utilizes existing indoor lighting infrastructure to provide high-precision indoor positioning services. However, due to the ...
详细信息
ISBN:
(数字)9798350368888
ISBN:
(纸本)9798350368895
The indoor imaging visible light positioning (VLP) technology based on the image sensor (IS) utilizes existing indoor lighting infrastructure to provide high-precision indoor positioning services. However, due to the high complexity of traditional imageprocessingalgorithms, existing VLP systems are unable to provide real-time positioning services. In response to the above issue, we propose an imageprocessing mechanism based on the inter-frame color feature matching method, which includes the equidistant sampling processing, low complexity decoding algorithm, and color feature matching-based LED region of interest (ROI) determination method. The results show that the system average positioning delay is 80.2ms, indicating better system real-time performance.
暂无评论