Generative adversarial network (GAN) model generates and discriminates images using an adversarial competitive strategy to generate high-quality images. The implementation of GAN in different fields is helpful for generating samples that are not easy to obtain. Image generation can help machine learning to balance data and improve the accuracy of the classifier. This paper introduces the principles of the GAN model and analyzes the advantages and disadvantages of improving GANs. The applications of GANs in image generation are analyzed. Finally, the problems of GANs in image generation are summarized.
The unified modeling language (UML) is used for the specification, visualization, and documentation of object-oriented software systems. Mobile UML (M-UML) is an extension of UML that considers mobility aspects, and a mobile statechart is an extension of the standard UML diagram that deals with the requirements for modeling, specifying, and visualizing mobile agent-based systems. However, mobile statecharts inherit UML’s lack of formal notation for analysis and verification purposes. The rewriting logic language Maude is a formal method that deals with mobile computations. In this paper, we propose a formalization of M-UML statechart diagrams using Maude to provide formal semantics for such diagrams. The generated Maude specifications are then used to analyze and check the systems using Maude analytical tools. This approach is illustrated through an example.
Obtaining useful information from ambiguous information is a necessity in various fields. Ambiguous information can be handled quantitatively by using fuzzy theory, and representing it in an easy-to-understand manner is critical. One solution is to visualize an ambiguous relationship by using fuzzy graph representation, which has the essential characteristic of expressing variable relationships in between its nodes. We previously proposed an algorithm to draw intelligible and comprehensive fuzzy graphs. This study describes an improved drawing method for that graph drawing algorithm. As a result, highly related nodes were arranged closer to one another, and the display area was reduced. This method can be used as an effective means of expressing the results of ambiguous information analysis.
A novel culture-based multiswarm artificial bee colony (CMABC) algorithm was proposed to address dynamic optimization problems. The historical experience of sub-swarms is preserved as cultural knowledge to guide the subsequent evolutionary process. Experiments were conducted on the moving peaks benchmark function. The results show that the CMABC algorithm was better than, or at least comparable to, the basic ABC algorithm, and other state-of-the-art algorithms.
This study proposed a quantitative evaluation method for vitreous opacities using motion video. The proposed method focused on moving turbidity in the vitreous. The moving turbidity appeared as an inter-frame difference, which was calculated from two consecutive frames. Therefore, the degree of vitreous opacity was estimated using this inter-frame difference. The proposed method was applied in the experiments to actual motion videos obtained using slit-lamp examination. The effectiveness of method was confirmed using the t-test and linear discriminant method.
The landmark project RoboCup is a well-known international robotics challenge that aims to advance robotics and AI research, with the end goal of developing robots capable of playing a game of soccer autonomously. Self-localization is one of the important elements for an autonomous soccer playing robot because the position information of the robot becomes a determinant of strategic behavior and cooperative operation. Although local searching is accurate, the lack of global searching results in the kidnapped robot problem. Thus, we propose a self-localization method that generates the searching space based on model-based matching using information regarding the white lines on the soccer field. The robot’s position is recognized by optimizing the fitness function using a genetic algorithm (GA). In this report, we adjust the parameter set of the GA on the basis of preliminary experiments and evaluate the accuracy of the proposed self-localization method. We verified that the proposed method enables real-time reversion to correct the position from the kidnapped position using the global/local searching ability of the GA.
Approximately 600,000 to 1,000,000 patients are diagnosed with rheumatoid arthritis (RA) in Japan. To provide appropriate treatment, it is necessary to accurately measure the progression of RA by diagnosing the disease several times a year. The modified total sharp score (mTSS) calculated from hand X-ray images is a standard diagnostic method for RA progression. However, this diagnostic method is time-consuming as the scores are rated at as many as 16 points per hand. Accordingly, in order to shorten the diagnosis time of RA patients and improve the quality of diagnosis, the development of computer-aided diagnosis (CAD) systems is expected. We have previously proposed a CAD system that can detect finger joint positions using a support vector machine and can estimate the mTSS using ridge regression. In this study, we propose a fully automatic detection method of RA score evaluation points in the carpal site from simple hand X-ray images using deep learning. The proposed method first segments the carpal site using deep learning. Next, the RA evaluation points are automatically determined from each segment based on prior knowledge. Experimental results on X-ray images of the hands of 140 patients with RA showed that the mTSS evaluation point at the carpal site could be detected with an average error of 25 pixels. This study enables the automatic detection of RA score evaluation points in the carpal site. In the diagnosis of RA, the time required for diagnosis can be reduced by automating the determination of diagnostic points by physician.
The mechanical arm is an important component in many types of robots; however, in certain production lines, the conventional grasp strategy cannot satisfy the demands of modern production because of several interference factors such as vibration, noise, and light pollution. This paper proposes a new grasping method for manipulators in stamping automatic production lines. Considering the factors that affect grasping in the production environment, the deep deterministic policy gradient (DDPG) method is selected in this study as the basic reinforcement-learning algorithm, and this algorithm is used to grasp moving objects in stamping automatic production lines. Owing to the low success rate of the conventional DDPG algorithm, the hindsight experience replay (HER) is used to improve the sample utilization efficiency of the agent and learn more effective tracking strategies. Simulation results show an 82% mean success rate of the optimized DDPG-HER algorithm, which is 31% better than that of the conventional DDPG algorithm. This method provides ideas for the research and design of the sorting system used in stamping automation production lines.
In this study, the morphology of the PPG signal has been analyzed to be a potential cardiovascular marker for physiological stress. The morphology of the PPG signal was quantified as signal quality index by comparing the template beat (extracted from resting conditions) to the PPG beats recorded during vigorous physical activity. Data was taken from eight subjects where they performed some physical activities ranging from low to high intensity. It was found that, the mean and standard deviation of correlation coefficient between non-stress condition template beat and annotated PPG beat, 89.43±5.17 (%) and 44.23±10.48 (%) for non-stress and stress beat respectively with P value of 2.04*10-06 shows significantly difference between correlation coefficients (stress and non-stress). Whereas, mean and standard deviation of dynamic time warping correlation coefficients are 93.43±5.06 (%) and 85.93±4.18 (%) for non-stress and stress beat respectively with P value of .04. The morphology results corroborate the findings from the traditional HRV parameters generally used for stratifying stress.
The sense of sleepiness and fatigue that occurs at around 2 p.m. is known as the “post-lunch dip (PLD).” It causes a transient decline in brain function, including cognitive function, attentiveness, and arousal level. Various research hypotheses have been proposed for the mechanism of occurrence of PLD, including explanations involving blood sugar spikes or the inhibition of neuropeptides. However, the evidence for these hypotheses is poorly constructed, and none of them is widely recognized as an explanation for the mechanism. The establishment of quantitative evaluation indicators for the decline in brain function caused by PLD is essential to clarify the mechanism of occurrence of PLD. In this study, a demonstration experiment was conducted focusing on P300 and contingent negative variation (CNV), which are types of event-related potentials (ERP), as evaluation indicators of PLD. The subjects were 14 healthy young people, and the meal load used was two slices of white bread and 285 mL of water. In the experiment, measurements were taken four times in total (preprandial, immediately postprandial, 40 min postprandial, and 80 min postprandial). The Stanford sleepiness scale (SSS) and a subjective questionnaire about fatigue using a visual analog scale (VAS) were administered before each measurement. The results confirmed that, at 40 min postprandial, when a significant increase in SSS was observed, there was a reduction in the area of early CNV and late CNV and a prolongation of P300 latency (p<0.05). An evaluation using late CNV also confirmed a reduction in area immediately postprandial that could not be confirmed in the SSS.
There is presently a shortage of nurses in Japan, with a further shortage of 3,000–130,000 nurses expected. There is also shortage of scrub nurses. Scrub nurses are nurses who work in the operating room. The main job of scrub nurses is to assist surgeons. Scrub nurses are a high turnover rate, because it is a difficult job. Therefore, system for assisting scrub nurses are needed. The purpose of this study was to develop a robotic scrub nurse. As a first step, a detection system for surgical instruments was developed using the “Faster Region-Based Convolutional Neural Network” (Faster R-CNN). In experiments, computer graphics (CG) model images and 3D-printed model images were evaluated, and the system showed high accuracy. Consequently, the Faster R-CNN system can be considered as suitable for detecting surgical instruments.
In the process of target detection with active light sources as calibration objects, air scattering and air absorption cause a significant loss of light energy, resulting in distortion and fragmentation of the spot shape. Inspired by band-pass filtering, this study proposes a target detection method based on variable frame rate sampling of an active light source. It primarily adopts i) image modulation for collecting the active light source signal with a specified frequency and subtracting the background, and ii) variable frame rate sampling for further weighted average to attenuate the dynamic noise. The experimental results show that the proposed method can efficiently eliminate static background, suppress dynamic noise, and detect the target location without illumination and background requirements.
Different imaging conditions often result in different color reproductions. Hence, color reproductions must be calibrated when images are captured under different imaging conditions. Herein, a new color calibration method based on iterative distributed transfer (IDT) is proposed. IDT is used to preliminarily calibrate color reproductions, and the results are known as preliminary results. Because IDT may result in unnatural colors, namely false colors, the projected color components of the input image are used to suppress the false colors, and the results are the final results. To obtain the final results, a series of projection coefficients must be calculated. By minimizing the objective function, in which the preliminary results are used as the target for the final results, the projection coefficients are calculated. Simultaneously, the weight associated with the false color is incorporated into the objective function such that the final results do not depend significantly on the preliminary results when the IDT yields false colors. Meanwhile, to quantitatively evaluate the false color, a false color index is proposed herein. The proposed method can suppress false colors and offers high color-adjustment capabilities. The effectiveness of the method is verified based on evaluation indexes and the observation of experimental results.
Indoor localization based on Bluetooth low energy (BLE) beacons has been rapidly developed, and many approaches have been developed to achieve higher estimation accuracy. In these methods, the received signal strength (RSS) is the input. However, the measurement of indoor environments is affected easily; the signal may be reflected and attenuated by obstacles such as the human body, walls, and furniture, which creates a challenge for methods based on signal mapping. In this study, BLE signal characteristics are investigated in an indoor localization setting. An experiment is performed using one BLE beacon and multiple receivers installed at different wall and ceiling positions. The raw RSS is observed, and the relationship between the BLE beacon signal strength characteristics against the human body effect as well as the receiver’s placement in the observation area are discussed. Signal mapping is performed, where the signal strength is measured from all receivers simultaneously. The position estimation accuracy is examined based on different data scenarios. The results show that the estimation position estimated by the BLE beacon based on extensive BLE beacon data does not affect the estimation accuracy.