Search

Dataset quality is an integral aspect of image object detection. Consequently, poor-quality datasets are often blamed for limiting model performance. Conventional convolutional neural network models attempt to address the model performance issue and have achieved high accuracy; however, these models fail to detect objects with photometric and geometric variations in similar images. To address this object detection issue, we adopt selective image augmentation by identifying potential weak points of photometric factors, such as brightness and sharpness. A stall-housed pig dataset and a food product dataset with complex and clear backgrounds, respectively, were selected to test our hypothesis. Images from both test datasets were investigated with photometric variations ranging between −60% $\sim ~60$ % with a step size of 20% to identify the weak points of the dataset. The test results demonstrate that a photometrically balanced dataset with selective augmentation significantly outperforms the original dataset in terms of high accuracy and robustness.

In this paper, a new pixel circuit for active matrix organic light-emitting diode (AMOLED) display that can achieve high uniformity in low gray levels and its driving method are proposed. The proposed circuit compensates for threshold voltage variation of thin-film-transistors (TFTs), with the structure that minimizes the loss of sensed threshold voltage. However, the high current error rate in extremely low gray level is unavoidable, as the driving TFT (DRT) operates in subthreshold region, where the current difference caused by the threshold voltage variation can be severe. To suppress high error rates in low gray levels, the operation region of DRT is restricted to the saturation region, by adopting duty ratio modulation (DRM) method. With the DRM method, low gray is expressed with high current value and short emission time. The viability of the proposed circuit and its operation are analyzed with HSPICE. Compared to the conventional driving method, DRM significantly reduces the current error rate in low gray area. The proposed circuit is fabricated within 220 $\mu {\mathrm {m}} \times 440 \mu {\mathrm {m}}$ . The measurement of the circuit also verified the capability of the proposed circuit and the DRM method.

This study aimed to evaluate the effectiveness of our newly developed virtual reality head-mounted display (VR-HMD) “walker avoidance” game in reducing step-aside reaction time (SART) and enhancing agility in collision avoidance. Fifteen young adults in experimental group (EG) engaged in the “walker avoidance” game, while another 15 young adults in the control group (CG) played the “first touch” tutorial. The results showed the EG had significant decreases (p < 0.01) in both SART-standing and SART-walking when compared with pre-intervention measurements. Compared with the CG, the EG SART-standing exhibited significant decreases in both the first (p = 0.001) and second (p < 0.001) measurements post-intervention; the EG SART-walking demonstrated significant decreases in all (p < 0.05) measurements, except for pre-intervention measurement. One-dimensional statistical parametric mapping (spm1d) also demonstrated significant differences in most of the electromyography and forefoot/hindfoot ground reaction force results because the step-aside movement became quicker in the EG following training. After pushing the leg-heel contact, the EG participants made a toe-off sooner than the CG participants. Following two sessions of our newly developed “walker avoidance” game, conducted 1 week apart, the EG exhibited less collisions with virtual pedestrians and reduced reaction times to unpredictable directional change measurements compared with the CG. This study demonstrated the effectiveness of this targeted VR training program in improving motor function, which introduced a novel approach to rehabilitation as a digital therapy. It offers innovative perspectives and an approach for clinical rehabilitation, while also providing new ideas for the VR content development industry.

In this paper, we proposed a new low-temperature polycrystalline oxide (LTPO) thin-film transistor (TFT) pixel circuit for micro light-emitting diode (μ LED) displays that produces a highly stable and uniform driving current. The proposed pixel circuit suppresses the current level change along with the sweep signal due to the parasitic capacitances and compensates for the TFT's threshold voltage (VTH) variation-induced current error, including even falling shape. In addition, the proposed circuit produces a constant current regardless of the data voltage. As a result, a relative current error rate of less than 2% was achieved across all gray levels under the ±0.5 V VTH fluctuation. The proposed circuit was verified using HSPICE with a low-temperature polycrystalline silicon (LTPS) TFT and amorphous indium-galliumzinc- oxide (a-IGZO) TFT model based on the measured data. The simulation analysis confirmed that the optimal sweep signal input position and pulse width modulation (PWM) and constant current generation (CCG) parts connecting method were key design points for stable and uniform performance.

With the development of artificial intelligence, there have been many attempts to incorporate artificial intelligence into algorithmic trading. In particular, reinforcement learning, which aims to solve dynamic decision-making problems, is attracting attention because of its high utilization in algorithmic trading. In this paper, we will implement a simple Deep Reinforcement Learning (DRL) trading robot to check the performance of DRL. In addition, we tried to find out how much performance improvement can be achieved by comparing a robot that learned a single stock data with a robot that learned stock data, market index, and interest rate data. This paper aims to develop a stock investment robot using a Proximal Policy Optimization (PPO) reinforcement learning algorithm and analyze the performance of the robot. The first robot used only the stock data of APPL INC, a single stock, as input, and the second robot used stock data of APPL INC and the S&P 500 index together with US interest rate data as input data. Afterward, the stock investment performance of the two robots for APPL INC was comparatively analyzed using the test data.

Rehabilitation training is essential for a successful recovery of upper extremity function after stroke. Training programs are typically conducted in hospitals or rehabilitation centers, supervised by specialized medical professionals. However, frequent visits to hospitals can be burdensome for stroke patients with limited mobility. We consider a self-administered rehabilitation system based on a mobile application in which patients can periodically upload videos of themselves performing reach-to-grasp tasks to receive recommendations for self-managed exercises or progress reports. Sensing equipment aside from cameras is typically unavailable in the home environment. A key contribution of our work is to propose a deep learning-based assessment model trained only with video data. As all patients carry out identical tasks, a fine-grained assessment of task execution is required. Our model addresses this difficulty by learning RGB and optical flow data in a complementary manner. The correlation between the RGB and optical flow data is captured by a novel module for modality fusion using cross-attention with Transformers. Experiments showed that our model achieved higher accuracy in movement assessment than existing methods for action recognition. Based on the assessment model, we developed a patient-centered, solution-based mobile application for upper extremity exercises for hemiplegia, which can recommend 57 exercises with three levels of difficulty. A prototype of our application was evaluated by potential end-users and achieved a good quality score on the Mobile Application Rating Scale (MARS).

In this paper, we propose a new active-matrix organic light-emitting diode (AMOLED) pixel circuit using amorphous indium-gallium-zinc oxide (a-IGZO) thin film transistors (TFTs). The proposed pixel circuit consists of seven TFTs and two capacitors, compensating for both threshold voltage (VTH) and mobility variations. The simulation results show that the proposed pixel circuit can successfully compensate for mobility variation. Also, the mobility compensation stage positively affects a significant ${\mathrm{ V}}_{\mathrm{ TH}}$ fluctuation. The proposed circuit is fabricated within an area of $19.95\,\,\boldsymbol{\mu }\text{m}\,\,\times 39.9\,\,\boldsymbol{\mu }\text{m}$ that can achieve 635 pixels per inch (PPI), and the experimental results show relatively consistent current levels even under severe TFT variations.

Recognizing traffic accident events in driving videos is a challenging task and has emerged as a crucial area of interest in autonomous driving applications in recent years. To ensure safe driving alongside human drivers and anticipation of their behaviors, methods to efficiently and accurately detect traffic accidents from a first-person viewpoint must be developed. This paper proposes a novel model, named the TempoLearn network, which leverages spatio-temporal learning to detect traffic accidents. The proposed approach incorporates temporal convolutions, given their effectiveness in identifying abnormalities, and a dilation factor for achieving large receptive fields. The TempoLearn network has two key components: accident localization, for predicting when the accident occurs in a video, and accident classification based on the localization results. To evaluate the performance of the proposed network, we conduct experiments using a traffic accident dashcam video benchmark dataset, i.e., the detection of traffic anomaly (DoTA) dataset, which is currently the largest and most complex traffic accident dataset. The proposed network achieves excellent performance on the DoTA dataset, and the accident localization score, measured in terms of AUC, is 16.5% higher than that of the existing state-of-the-art model. Moreover, we demonstrate the effectiveness of the TempoLearn network through experiments conducted on another benchmark dataset, i.e., the car crash dataset (CCD).

In this paper, we present a novel pulse width modulation (PWM) operational scheme for micro-light-emitting-diode ( $\mu$ LED) displays using inverter and dynamic SWEEP signal slope. In the proposed PWM operational scheme, the slope of the SWEEP signal steepens as the gray level decreases. This adjustable SWEEP signal slope further shortens the falling time compared to the inverter-only strategy, leading to an exceptional capability to express extremely low gray levels. We implement the proposed scheme with a low-temperature poly-Si oxide (LTPO) thin-film transistor (TFT) pixel circuit and verify the operation through HSPICE based on measured TFT data. The simulation results show that the proposed scheme can significantly shorten the falling time to under $\mu$ s level. The gray level of 14 G is accurately expressed without wavelength shift or pulse distortion, achieving a falling time of $\mu$ s. In addition, we also verify the compensation accuracy of the pixel circuit, and the results indicate that the proposed pixel circuit can reliably compensate for TFT variations.

Grade schoolers and teenagers wonder how tall they will be, as there is a tendency to prefer taller stature for many years. Child’s height growth is one of the continuous interests of the parents from the past to the present for many reasons, not only their children’s outer beauty but also health status of children. Pediatricians also want to make sure a child is growing as expected because the height growth of children is an important indicator for monitoring a child’s nutrition and diseases. In many previous studies, adult height prediction method using growth curves is used widely. Unfortunately, growth curves are based on longitudinal cohort studies which are very challenging to conduct. That’s why it is hard to find the related studies for certain ethnic group. In this study, we collected 2,687 Korean height data including parental heights and children’s heights by ourselves in the same format as Galton’s Height data at 1880s in the United Kingdom. Then, we focus on the influence of parental height on child’s height conducting various analysis comparing Galton’s and Korean height data. Especially, we find out the linearity of child’s height varies depending on the combination of each parental height through visualization analysis. Finally, we propose our method of deploying the best among various machine learning techniques according to the combination of parental height. The combination is based on distribution of each parental height. And it outperforms achieving RMSE under 3.5 compared to single machine learning models which cannot achieve RMSE even under 4.0. It will be a simple and good application for many of pediatricians and parents who care a lot about their children’s height growth.

A video see-through head mounted display (VSHMD) has a digital camera set attached to the front of the HMD, which is a promising personal display device bringing a user into a realistic mixed world on demand. However, VSHMDs have an inherent risk of diminishing visual performance owing to the limited field of view (FOV) and poor display resolution etc. We investigated influences of VSHMD on the peripheral visual search performance of wearers. Nineteen participants performed a covert visual feature search task. All participants performed the tasks in four FOV conditions: 15°, 30°, 45°, and 60° with three types of viewing devices including VSHMD. The results showed that the VSHMD deteriorated the user’s peripheral visual search performance worse than the HMD, whereas no deterioration found in the FOV restriction or the low resolution of stimuli. This result implies that low display sharpness and additional video signal chain of VSHMD cameras may induce diminished peripheral visual perception. Designers should consider this effect when developing any augmented reality or mixed reality system. Such investigations will raise an important guideline for the practical usage of VSHMD and tele-operation.

In this paper, we propose a novel amorphous In-Ga-Zn-oxide (a-IGZO) thin film transistor (TFT) pixel circuit for pulse width modulation (PWM) driving active matrix micro light-emitting diode $(\mu $ LED) displays. The proposed pixel circuit compensates for wavelength shifts according to gray level and threshold voltage (VTH) variation of TFTs with only two capacitors. This compact operation can be achieved by simultaneous compensation for VTH along with data input in the constant current generation (CCG) circuit and the PWM circuit, which consist the pixel. Since the VTH variation in TFTs is compensated, the proposed pixel circuit can supply a uniform and stable current to the $\mu $ LED throughout all gray levels. The compensation accuracy was verified by HSPICE with an oxide TFT model based on the measurement data. In addition, we analyzed the optimization method to improve the PWM driving by controlling the electrical properties of TFT, and verified this approach by simulation.

This study investigates the emotional responses to vehicle interior lighting by measuring physiological reactions and psychological judgments. Individually controllable LEDs are installed in the front area of the passenger car (Sonata DN8, Hyundai Motor Company) and generate light stimuli by varying the attributes of the LEDs for the user experiment. Forty participants volunteered in the experiment and opted to evaluate either blue (468 nm) and orange (589 nm) light prior to the main session. Each participant was presented with twelve lighting signatures made with six lighting behaviors applied to two types of placements. During the experiment, we collected the alpha and beta waves using an electroencephalography (EEG) and subjective assessments of the lighting with seven questions. The ratio of the beta waves was the highest in the Blink mode, followed by the Fade and Collide modes. We combined the EEG responses with subjective assessments of seven adjective pairs toward evidence-based guidance for designing user-centered light in automobiles.

Integration of communication and radar functions with a single waveform has been actively investigated in various wireless communication applications, including unmanned aerial vehicles (UAVs). This means that communication frames consisting of a pilot part and subsequent data part can be utilized to transmit information and detect the surrounding objects simultaneously. Because a predetermined waveform is required for the radar function, the pilot part in the communication frame can be utilized for radar purposes. Specifically, the pilot is used for both channel estimation, which is required for data decoding, and a radar waveform. Assuming that the length of the pilot and data parts is given and the transmit energy for each frame is limited, the optimal transmit power for the pilot and data parts can be analytically obtained by considering both radar and communication performance metrics. The optimality of the proposed analytical solution was verified through numerical simulations.

The aim of this research is to propose a binary segmentation algorithm to detect the change points in financial time-series based on the Iterative Cumulative Sum of Squares (ICSS). The proposed algorithm, entitled KW-ICSS, utilizes the non-parametric Kruskal-Wallis test in cross-validation procedures. In this regard, KW-ICSS can quickly detect the change points in non-normally distributed time-series with a small number of observations after the change points than the state-of-the-art ICSS algorithm, entitled AIT-ICSS. For the simulated financial time-series whose true location of the change point is known, KW-ICSS detects the change points with the average true positive rate of 81% for the different number of change points, whereas AIT-ICSS only exhibits 72.57%. Also, KW-ICSS’s mean absolute deviation between the true and detected change points is less than that of AIT-ICSS for different significance levels. The experiment also finds that the significance level, the model parameter, should be set to less than 10%. For the real-world financial time-series whose true location of change points is unknown, KW-ICSS’s robust detection of change points is observed from fewer detected change points and longer intervals between them. Furthermore, KW-ICSS’s trend prediction for the short-term future performs with an average of 92.47% accuracy, whereas AIT-ICSS shows 90.69%. Therefore, we claim that KW-ICSS successfully improves AIT-ICSS.

Cortical connectivity analysis is a widely used method for understanding the causes of neurological disorders and related brain mechanisms. Although there exist numerous activity analysis toolboxes for functional near-infrared spectroscopy (fNIRS), there are only a few cortical connectivity analysis toolboxes. In 2019, we released a MATALB toolbox named OptoNet, which has helped researchers to analyze brain networks using fNIRS. In this study, we developed an advanced MATLAB toolbox, named OptoNet II, to add new features that overcome the shortcomings of OptoNet. With these new features, OptoNet II can efficiently analyze cortical connectivity according to brain region using any fNIRS channel sets and can present the results of two connectivity analyses with auto-thresholding based on surrogate tests. To evaluate the efficacy of the new functions, the finger-tapping task experiment was carried out before and after transcranial direct current stimulation (tDCS) in the primary motor area. OptoNet II can efficiently show the effects of tDCS on functional brain region connectivity, which has been difficult to confirm by conventional methods. In this article, we propose the OptoNet II as a useful and efficient toolbox for researchers who want to perform cortical connectivity analysis using fNIRS.

This research examines and proposes an investment strategy by combining the natural language processing on the equity research reports published in the Korean financial market and machine learning algorithms for binary classification. At first, we deduce the part-of-speech from the report using the KoNLPy and Mecab. Then, we define 33 features as the input variables and perform the binary classification on the price direction of the stocks recommended in the report using various machine learning algorithms. Note that we investigate the model performance in detail by dividing the entire period into three sub-periods, including pre-COVID-19 for the sideways market, COVID-19 for the crashing market, and post-COVID-19 for the extreme bullish market. We confirm that the random forest is the best classifier for all periods, so we utilize its results on positively predicted stocks in the test set as the investment universe for the monthly re-balancing and buy-and-hold investment. The proposed strategy shows a significantly higher return on investment than benchmarks during the pre-COVID-19 and COVID-19 periods, whereas the comparable return during the post-COVID-19.

This study proposes a new assistive device for the blind that uses more than one-dimensional data to draw objects. The study aims to convert three-dimensional (3-D) spatial information into sound information using 6-axis and ultrasonic sensors, and to draw a 3-D depiction of the space ahead for the user. Fourteen participants were involved in testing, wherein 4 were visually impaired. Moreover, the male to female ratio was 7:3, with the average age of participants at 28.8 years. An initial sound recognition experiment was designed to assess the device's accuracy through participant use. Recognition rates were 70% for normal participants and 88% for the blind participants. Additional experiments expanded the environmental conditions by requiring participants to discern the distances of 10 objects, positioned at both high and low locations. Two different scenarios were employed: stationary and walking scenarios. The stationary distance measurement participants scored an average of 96 points, while the walking participants averaged 81 points. Under the given conditions, this study found that its assistive device for the visually impaired can draw a 3-D space with 88.5% accuracy. This probability promises a basic level of utility that can assist those with visual impairment in controlled environments, such as hospitals and homes.

This study proposes FingerTouch, a method of touch interaction using a head-mounted display for mobile augmented reality. FingerTouch allows users to manipulate virtual content with one-finger touch interaction regardless of the material or tilt of the plane the finger is touching. In addition, users can interact freely with virtual contents using FingerTouch. As the prototype developed in this study uses only one inertial measurement unit sensor attached to a fingernail, it features high mobility and allows users to feel natural tactile feedback, which is important for performing everyday tasks. The user evaluation of FingerTouch indicated that it provides high accuracy, with an average cursor navigation error of 2.15 mm and an average finger gesture recognition accuracy of 95% across 22 participants, two surface orientations, and three surface materials.

LiDAR sensors have the advantage of being able to generate high-resolution imaging quickly during both day and night; however, their performance is severely limited in adverse weather conditions such as snow, rain, and dense fog. Consequently, many researchers are actively working to overcome these limitations by applying sensor fusion with radar and optical cameras to LiDAR. While studies on the denoising of point clouds acquired by LiDAR in adverse weather have been conducted recently, the results are still insufficient for application to autonomous vehicles because of speed and accuracy performance limitations. Therefore, we propose a new intensity-based filter that differs from the existing distance-based filter, which limits the speed. The proposed method showed overwhelming performance advantages in terms of both speed and accuracy by removing only snow particles while leaving important environmental features. The intensity criteria for snow removal were derived based on an analysis of the properties of laser light and snow particles.

Image processing systems are widespread with the digital transformation of artificial intelligence. Many researchers developed and tested several image classification models using machine learning and statistical techniques. Nevertheless, the current research seldom focuses on the quality assurance of these models. The existing methods lack to verify the quality assurance, with the lack of test cases to prepare the evaluation dataset to test the model, which can cause critical drawbacks in the nuclear field and defense system. In this article, we discuss and suggest the preparation of the evaluation dataset using improved test cases through Cause-Effect Graphing. The proposed method can generate the evaluation dataset with automated test cases through the quantification method, which consists of 1) image characteristic selection 2) creating the Cause-Effect graphing approach of the image with the feature, and 3) generate all possible test coverage. The testing is performed with the COCO dataset, which shows the declining prediction accuracy with the adjusted brightness and sharpness ranging between -75 to 75%, which indicates the negligence of the important characteristics in the existing test dataset. The experiment shows the prediction fails while sharpness is less than the 0%, and the brightness fails at -75% with less number of detection object between -50% and 75%. This indicates that characteristic changes affects the prediction accuracy and the number of detected objects in an image. Our approach proves the importance of the characteristic selection process for the overall image to generate a more efficient model and increase the accuracy of object detection.