首页期刊导航|International journal of innovative computing, information and control
期刊信息/Journal information
International journal of innovative computing, information and control
ICIC International
ICIC International
月刊
1349-4198
International journal of innovative computing, information and control/Journal International journal of innovative computing, information and controlISTPAHCIESCI
查看更多>>摘要：This research presents an automated system for detecting tuberculosis bacteria in sputum samples using deep learning. The study primarily focuses on developing a detection algorithm based on the Y0L0v8 architecture to accurately identify tuberculosis bacteria. The system integrates a computer, camera, and microcontroller to automate the movement of preparation and focus servos for enhanced scanning efficiency. Performance testing of the microscope-robot system reveals a precision rate of 81.28% and a recall of 90.71%, demonstrating a balanced performance. The precision-confidence curve shows high confidence in classifying positive samples, while the recall-confidence analysis indicates the model's strong ability to identify true positives. Despite the 75.1% overall accuracy, there are some false positives and false negatives. The system holds promise for accelerating and improving tuberculosis detection. Future work will focus on refining the model to enhance accuracy and overcome existing limitations. This research contributes to the development of faster, more reliable diagnostic technology for tuberculosis detection in medical settings.
DAO THANH HUYENNGHIA THI MAISUZUKA TANAKAKOTARO HASHIKURA...
287-305页
查看更多>>摘要：Robots capable of human communication play crucial roles in various daily tasks, particularly in nursing and mental healthcare. Effective emotion expression is essential for robots to perform these tasks proficiently. However, prior methods for robot emotion expression often involve complex designs, making them costly and impractical for widespread application. Additionally, color-based methods for emotional expression have not been thoroughly explored within the Vietnamese cultural context, where nuances in color interpretation are significant. This study is the first to specifically investigate cultural differences in emotional expression for robots using colors and gradients tailored to Vietnamese demographics. Moreover, we propose a simple, cost-effective design for robot emotion expression that addresses the shortcomings of previous approaches, which rely on complex and expensive designs, while still being capable of expressing complicated emotions such as dyads. Experimental results demonstrate the effectiveness of these models in conveying extended emotional states and dyads, independent of traditional human-like expressions. These findings underscore the practical potential of such models, particularly within Vietnamese nursing and mental healthcare sectors.
查看更多>>摘要：The variation trend of energy ridge in the noise spectrum of mine hoist is an important index of its health state. Due to the interference of background noise, the energy ridges in the hoist spectrum have unclear boundaries and poor continuity, making them hard to separate and analyze. To address this challenge, we propose a new end-to-end training deep neural network that extracts and fuses multi-scale features of spectrographs using multiple encoder-attention module-decoder blocks. The multi-scale features enable the network to capture the energy ridges in the spectrum more accurately and robustly. The small-scale feature map is used for coarse localization, and the large-scale feature map is used for fine refinement. We design and train our network, named DeepRidge, on a noise spectrum dataset of mining hoist collected from a preliminary experiment. We compared DeepRidge with other state-of-the-art methods, and the test results show that DeepRidge achieves better accuracy at smaller parameter sizes, with an average precision (AP) of 0.826. We also conduct experiments to find the optimal network configuration for the energy ridge detection task in the hoist noise spectrum.
查看更多>>摘要：Remote sensing images are often affected by atmospheric factors such as haze during the acquisition process, resulting in blurring and low contrast in the collected remote sensing images. This problem impacts the image quality, thereby affecting the analysis of remote sensing images. To mitigate the impact of haze on remote sensing images, a generative adversarial network is proposed. It comprises a generative network and an adversarial network. Firstly, a novel feature extraction module is designed to enhance the capability of extracting useful information from remote sensing images. It enables the network to focus more on regions with dense haze, allowing it to extract more important information while filtering out irrelevant details. Secondly, a residual attention module is designed which can allocate different weights based on varying haze density in the feature map. This module readjusts features outputted by the encoder, facilitating better image restoration. Thirdly, a multi-scale module is also incorporated to extract feature information across various image scales. Lastly, a color feature extraction module is designed to extract color features. The novel feature extraction module, residual attention module, multi-scale module, and color feature extraction module are utilized for constructing the generative network. Besides, an adversarial network is also designed to indirectly enhance the dehazing capability of the generative network. Synthetic and real datasets are used to test six different methods for dehazing remote sensing images, respectively. The proposed method achieves higher PSNR, SSIM, and lower MSE on the synthetic remote sensing dataset. On the other hand, it achieves lower PIQE, BRISQUE, and higher MetaIQA on the real remote sensing dataset. The proposed method has best performance in dehazing remote sensing images than other methods.
查看更多>>摘要：Zadeh introduced the concept of fuzzy sets in 1965. In 1986, Atanassov introduced intuitionistic fuzzy sets, a generalization of fuzzy sets. Pythagorean fuzzy sets are a recent extension of intuitionistic fuzzy sets introduced by Yager. The aim of this study is to apply the concept of Pythagorean fuzzy sets to IUP-algebras and introduce the notions of Pythagorean fuzzy IUP-subalgebras, Pythagorean fuzzy IUP-ideals, Pythagorean fuzzy IUP-filters, and Pythagorean fuzzy strong IUP-ideals. The study identified a relationship between four concepts, showing that Pythagorean fuzzy IUP-ideals and Pythagorean fuzzy IUP-subalgebras are generalizations of Pythagorean fuzzy strong IUP-ideals in IUP-algebras, where the latter can only be a constant Pythagorean fuzzy set. Additionally, Pythagorean fuzzy IUP-filters were found to be a further generalization of Pythagorean fuzzy IUP-ideals and IUP-subalgebras. Their properties are investigated, and the characteristic Pythagorean fuzzy sets, the upper t- (strong) level subsets, and the lower t-(strong) level subsets of the Pythagorean fuzzy set are studied.
查看更多>>摘要：Reliable encrypted traffic classification is fundamental for advancing cyber-security and effectively managing exponentially growing data streams. The success of large language models in fields such as natural language processing demonstrates the feasibility of learning general paradigms from extensive corpora, making pre-trained encrypted traffic classification methods a preferred choice. However, attention-based pre-trained classification methods face two key constraints: the large number of neural parameters is unsuitable for low-computation environments like mobile devices and real-time classification scenarios, and there is a tendency to fall into local minima, leading to overfitting. We develop a shallow, lightweight Transformer model named LETformer. We utilize sharpness-aware optimization during pre-training to avoid local minima while capturing temporal features with relative positional embeddings and optimizing the classifier to maintain classification accuracy for downstream tasks. We evaluate our method on four datasets - USTC-TFC2016, ISCX-VPN2016, ICCXTOR, CICIOT2022. Despite having only 17.6 million parameters, LETformer achieves classification metrics comparable to those of methods with ten times the number of parameters.
查看更多>>摘要：During this era of data science, vast data are collected to infer the unknown plant and unknown controller, existing in a closed loop system structure, which corresponds to direct data driven modelling and direct data driven control. This new paper studies the latter direct data driven control, i.e., designing that feedback controller based on measured input-output data sequence directly. Firstly, after introducing the basic essence of direct data driven control, nonparametric controller and parameterized control are all derived through our own derivations. Secondly, due to the widely application of adaptation into other aircraft control, adaptive idea is applied to identifying the unknown controller parameters, and its related algorithm is formulated in detail while analyzing the termination condition. Thirdly, to guarantee the dual goals, i.e., perfect tracking and asymptotic unbias estimation, Lyapunov function is constructed to derive the accurate nonparametric controller and parameterized controller. Fourthly, application of our considered adaptive direct data driven control is proved in a practical example of aircraft control. Generally, the mission of this paper is to apply adaptive idea into direct data driven control although research only on direct data driven control is very mature. Furthermore, perfect tracking and asymptotic unbias controller are guaranteed through Lyapunov function analysis.
查看更多>>摘要：In this paper, the diagnosis and location issue of open-circuit faults in neutral-point-clamped three-level inverters is analyzed. A hybrid method based on model-based and data-driven fault diagnosis is proposed for the insulated gate bipolar transistors open-circuit faults. First, the three-phase current residual is attained by subtracting the true value of the three-phase current generated by the inverter with the estimated three-phase current generated by the state estimator. Second, the Park's vector modulus and wavelet transformation algorithms are utilized for normalization of three-phase current residuals and feature analysis. Then, the 11 fault features of current residuals are extracted and the datasets of fault feature are established. Finally, the fault feature samples are utilized to train the random forest model to achieve state classification. The proposed method can improve the diagnostic accuracy compared with traditional fault diagnosis methods. The effectiveness and the robustness of this method under various conditions are validated by experimental results.
查看更多>>摘要：This paper presents a model to determine the optimal surface of a hollow circular footing or annular strip footing (the width and the position where pressure is zero, since the radius is subject to the conditions of the superstructure), assuming that the soil is elastic, the soil pressure distribution is linear and the surface in contact with the soil works partially in compression. Some authors show the optimal surface of solid circular footings taking it into account that the area in contact with the soil works partially under compression, and other present annular strip footings or hollow circular footings assuming uniform soil pressure. The formulation is developed by integration to determine the axial load "P" and a resultant moment "Mr " from the moments "M_x " and "M_y ". Six numerical examples are presented to determine the soil contact area for hollow circular footings under an axial load and a resultant moment. A comparison is also made with the results of other authors and the results show that using the proposed model savings of up to 43.75% can be achieved.
查看更多>>摘要：ABSTRACT. Multiple clock sources are needed in mobile communication devices to drive each system. As a clock source, the authors have previously proposed a frequency multiplier using a double-edge counter and demonstrated its effectiveness. This paper proposes an m + n/k multiplier based on a multi-phase clock. This circuit can realize m + n/k multiplication at non-integer frequencies of the input signal by employing the 1 + n/k divider, proposed by the authors, as the basis for the multi-phase clock counting circuit. The characteristics were verified through simulations using Verilog-HDL. It was clarified that the steady-state frequency error of the output signal relative to the input signal frequency corresponds to a one-phase difference of the multi-phase clock. Furthermore, it was confirmed that a regular multiplied signal is obtained within two cycles of the input signal.
NETL
NSTL