• Journal of Advanced Navigation Technology
• /
• v.23 no.1
• /
• pp.77-83
• /
• 2019

In this paper, we propose an object tracking algorithm based on color histogram and convolutional neural network model. In order to increase the tracking accuracy, we synthesize generic object tracking using regression network algorithm which is one of the convolutional neural network model-based tracking algorithms and a mean-shift tracking algorithm which is a color histogram-based algorithm. Both algorithms are classified through support vector machine and designed to select an algorithm with higher tracking accuracy. The mean-shift tracking algorithm tends to move the bounding box to a large range when the object tracking fails, thus we improve the accuracy by limiting the movement distance of the bounding box. Also, we improve the performance by initializing the tracking start positions of the two algorithms based on the average brightness and the histogram similarity. As a result, the overall accuracy of the proposed algorithm is 1.6% better than the existing generic object tracking using regression network algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.202-205
• /
• 2005

The purpose of this work is to identify the characteristic of eye movement for visual training of the elderly. This investigation is to examine the relationship between the head and the eye motor system for the localization of visual target direction in three-dimensional space. All experiments were performed in a soundproof chamber. Twenty-one red LEDs (Light-Emitting Diode, Brightness: 20cd/$m^2$) arrayed in three lines on a half circle-surrounding panel are used. LEDs are horizontally 30 degree apart and vertically 20 degree apart from each other. The condition of stimulation is random and anti-saccade. Physiological parameter such as EOG (Electro-Oculography) was measured by BIOPAC system. We measure the mean latency time, which is the time from the start of visual stimulation to the response of the human body. The result shows that the mean latency time is short in the case of the condition of anti-saccade, the fixed head and a quarter visual stimulation. This finding can be used in developing programs for various visual improvements for the elderly by analyzing the characteristic of eye movement.

• Journal of Information Processing Systems
• /
• v.12 no.3
• /
• pp.436-454
• /
• 2016

This research presents the battery discharge rate models for the energy consumption of mobile phone batteries based on machine learning by taking into account three usage patterns of the phone: the standby state, video playing, and web browsing. We present the experimental design methodology for collecting data, preprocessing, model construction, and parameter selections. The data is collected based on the HTC One X hardware platform. We considered various setting factors, such as Bluetooth, brightness, 3G, GPS, Wi-Fi, and Sync. The battery levels for each possible state vector were measured, and then we constructed the battery prediction model using different regression functions based on the collected data. The accuracy of the constructed models using the multi-layer perceptron (MLP) and the support vector machine (SVM) were compared using varying kernel functions. Various parameters for MLP and SVM were considered. The measurement of prediction efficiency was done by the mean absolute error (MAE) and the root mean squared error (RMSE). The experiments showed that the MLP with linear regression performs well overall, while the SVM with the polynomial kernel function based on the linear regression gives a low MAE and RMSE. As a result, we were able to demonstrate how to apply the derived model to predict the remaining battery charge.

• The Bulletin of The Korean Astronomical Society
• /
• v.38 no.1
• /
• pp.58.2-58.2
• /
• 2013

We present a photometric study of the globular clusters (GCs) in the giant elliptical galaxy M86 in the Virgo Cluster, using the Washington $CT_1$ images taken at the KPNO 4 m telescope. The color distribution of the GCs in M86 is bimodal. The radial number density profile of the blue GCs decreases more slowly as the galactocentric distance increases than that of the red GCs. The density profile of the red GCs is similar to the surface brightness profile of M86 stellar halo. The blue GCs have a roughly circular spatial distribution, while the red GCs have a spatial distribution somewhat elongated, which is consistent with the distribution of the galaxy stellar light. M86 GCs have the negative radial color gradient because the number ratio of the blue GCs to the red GCs increases as galactocentric radius increases. The mean color of the red GCs is similar to that of the stellar halo. The bright blue GCs in the outer region of M86 reveal a blue tilt that the mean colors of the blue GCs get redder as they get brighter. We discuss these results in comparison with other giant elliptical galaxies in the Virgo Cluster.

• Korean Journal of Remote Sensing
• /
• v.32 no.6
• /
• pp.589-601
• /
• 2016

Sea surface currents were estimated by applying the Maximum Cross Correlation (MCC), Zero-mean Sum of Absolute Distances (ZSAD), and Zero-mean Sum of Squared Distances (ZSSD) algorithms to Himawari-8/Advanced Himawari Imager (AHI) thermal infrared channel data, and the comparative analysis was performed between the results of these algorithms. The sea surface currents of the Kuroshio Current region that were retrieved using each algorithm showed similar results. The ratio of errors to the total number of estimated surface current vectors had little difference according to the algorithms, and the time required for sea surface current calculation was reduced by 24% and 18%, relative to the MCC algorithm, for the ZSAD and ZSSD algorithms, respectively. The estimated surface currents were validated against those from satellite-tracked surface drifter and altimeter data, and the accuracy evaluation of these algorithms showed results within similar ranges. In addition, the accuracy was affected by the magnitude of brightness temperature gradients and the time interval between satellite image data.

• Plant Resources
• /
• v.8 no.3
• /
• pp.235-243
• /
• 2005

The purpose of this research is to establish basic database on Cornus kousa by way of investigation, analysis and comparison of characteristics of leaf and flower, then fostering good cultivar on each morphological characteristics - leaf length, leaf width, petiole length, left lateral vein, right lateral vein. Leaf length and leaf width in Mt. Jiri shows big tendency which is 83.5 mm, 52.4 mm each in comparison with the whole mean of 72.5 mm, 41.2 mm, whereas populations of Suwon and Mt. Halla has leaf length of 66.0 mm, 65.7 mm - 9.8%, 10.4% lower value in comparison with seven mean population; leaf width is 38.4 mm, 35.3 mm - 7.3%, 16.7% lower than whole mean and shows lowest tendency among seven selected populations. Long width of flower and short width of flower in Boeun shows big tendency which is 99.9mm, 96.5 mm each in comparison with the whole mean of 76.0 mm, 73.6 mm, whereas populations of Mt. Halla has 50.1 mm, 48.2 mm which shows lowest tendency. On petal length, petal width and length of flower petiole, Boeun populations have bigger and Mt. Halla shows little tendency. The measurement result of flower colors on each population by using Spectrum Color Mater shows followings: populations of Mt. Duckyoo and Mt. Halla shows lower lightness than any other populations, but wholly shows higher lightness which is refers brightness in seven selected populations. Thus it shows peculiarities of white flower color.

• Korean Journal of Materials Research
• /
• v.12 no.10
• /
• pp.771-775
• /
• 2002

$Gd_2$$O_3$:Eu phosphor particles with nano-sized and non-aggregation characteristics were prepared by spray pyrolysis using the spray solution containing polymeric precursor and $Li_2$$CO_3$ flux material. Nano-sized $Gd_2$$O_3$:Eu phosphor particles had higher brightness than the commercial $Y_2$$O_3$:Eu phosphor particles. The $Gd_2$$O_3$:Eu phosphor particles had nano-size and non-aggregation characteristics after heat-treatment at $1000^{\circ}C$ when the addition amount of $Li_2$$CO_3$ flux was 1 wt.% and 3 wt.%. The mean size of particles were 200 nm and 400 nm when the amount of flux was 1 wt.% and 3 wt.%, respectively. The prepared phosphor particles had higher photoluminescence intensity than that of the commercial product regardless of the content of$ Li_2$$CO_3$ flux and had the maximum brightness when the content of flux was 5 wt %. The photoluminescence intensity of the nano-sized $Gd_2$$O_3$:Eu phosphor particles containing 3 wt.% $Li_2$$CO_3$ flux was 125% in comparison with that of the micron-sized $Y_2$$O_3$:Eu commercial product.

• Journal of Korea Multimedia Society
• /
• v.15 no.4
• /
• pp.439-448
• /
• 2012

TFT-LCD panel images have non-uniform brightness, noise signal and defect signal. It is hard to divide defect signal because of non-uniform brightness and noise signal, so various divide methods have being developed. In this paper, we suggest method to divide defective regions on TFT-LCD panel image by estimating a menas of two different size of windows, which is suggested by Eikvil et al., and using difference of them. But in this method, the size of detectable defects is restricted by the size of window, hence it has inefficient problem that the size of window have to increase to divide a large defect region. To solve this problem we suggest an algorithm which can divide various size of defects, by using Multi-scale and restrict a detectable size of defects in each scale. To prove an efficiency of suggested algorithm, we show that resulting images of real TFT-LCD panel images and an artificial image with various defects.

• KIISE Transactions on Computing Practices
• /
• v.20 no.12
• /
• pp.646-651
• /
• 2014

Histogram equalization generally has the disadvantage that if the distribution of the gray level of an image is concentrated in one place, then the range of the gray level in the output image is excessively expanded, which then produces a visually unnatural result. However, a gamma transformation can reduce such unnatural appearances since it operates under a nonlinear regime. Therefore, this paper proposes a new histogram equalization method that can improve image quality by using a gamma transformation. The proposed method 1) derives the proper form of the gamma transformation by using the average brightness of the input image, 2) linearly combines the earlier gamma transformation with a CDF (Cumulative Distribution Function) for the image in order to obtain a new CDF, and 3) to finally perform histogram equalization by using the new CDF. The experimental results show that relative to existing methods, the proposed method provides good performance in terms of quantitative measures, such as entropy, UIQ, SSIM, etc., and it also naturally enhances the image quality in visual perspective as well.

• Journal of Korea Multimedia Society
• /
• v.15 no.6
• /
• pp.748-760
• /
• 2012

In case brightness distribution is concentrated in a region, it is difficult to classify the image features. To solve this problem, we apply global histogram equalization and local histogram equalization to images. In case of global histogram equalization, it can be too bright or dark because it doesn't consider the density of brightness distribution. Thus, it is difficult to enhance the local contrast in the images. In case of local histogram equalization, it can produce unexpected blocks in the images. In order to enhance the contrast in the images, this paper proposes a local histogram equalization based on the Gaussian Mixture Models(GMMs) in regions of histogram. Mean and variance parameters in each regions is updated EM-algorithm repeatedly and then ranges of equalization on each regions. The experimental results performed with image of various contrasts show that the proposed algorithm is better than the global histogram equalization.