• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.1
• /
• pp.1-12
• /
• 2008

In this paper, we present a real-time algorithm for recognizing the vehicle color from the indoor and outdoor vehicle images based on GPU (Graphics Processing Unit) acceleration. In the preprocessing step, we construct feature victors from the sample vehicle images with different colors. Then, we combine the feature vectors for each color and store them as a reference texture that would be used in the GPU. Given an input vehicle image, the CPU constructs its feature Hector, and then the GPU compares it with the sample feature vectors in the reference texture. The similarities between the input feature vector and the sample feature vectors for each color are measured, and then the result is transferred to the CPU to recognize the vehicle color. The output colors are categorized into seven colors that include three achromatic colors: black, silver, and white and four chromatic colors: red, yellow, blue, and green. We construct feature vectors by using the histograms which consist of hue-saturation pairs and hue-intensity pairs. The weight factor is given to the saturation values. Our algorithm shows 94.67% of successful color recognition rate, by using a large number of sample images captured in various environments, by generating feature vectors that distinguish different colors, and by utilizing an appropriate likelihood function. We also accelerate the speed of color recognition by utilizing the parallel computation functionality in the GPU. In the experiments, we constructed a reference texture from 7,168 sample images, where 1,024 images were used for each color. The average time for generating a feature vector is 0.509ms for the $150{\times}113$ resolution image. After the feature vector is constructed, the execution time for GPU-based color recognition is 2.316ms in average, and this is 5.47 times faster than the case when the algorithm is executed in the CPU. Our experiments were limited to the vehicle images only, but our algorithm can be extended to the input images of the general objects.

• Journal of dental hygiene science
• /
• v.9 no.1
• /
• pp.121-127
• /
• 2009

A comparative analysis was conducted through this study on tooth brushing and a use of auxiliary oral hygiene devices for oral hygiene for 287 S Health College students some of whom major dental hygiene and some of whom don't relate to the department, and then has drawn out the results as below. 1. Relating to the frequency of tooth brushing, 2-3 times a day was the utmost case in both departments(p < 0.05). 2. In use of toothbrush bristle, the students of dental hygiene department used medium bristle(51.2%), while the student of other department used soft bristle(51.5%)(p > 0.05). 3. In the term of tooth brush used, 2~5 months was the utmost case in both departments(p > 0.05). 4. In tooth- brushing method, all the student of both departments said they knew how to do it(p > 0.05). 5. In the acquisition of tooth brushing methods, the students of dental hygiene acquired it from dentistry 44.4%, and the students of other department from TV or Radio 47.5% (p > 0.05)which was a quite high rate. 6. In reference to recognition level on auxiliary oral hygiene devices, dental hygiene students said Yes 79.8%, and the other side students said No 63.9%(p < 0.05). In reuse of auxiliary oral hygiene devices, a great number of students of both departments said No (p < 0.05). 8. In reference to the chance to use auxiliary oral hygiene devices, 58.1% of dental hygiene students chose them with their own judgement, and 37.8% of other department students recommended by dentist's offices(p < 0.05). 9. Relating to the reason for not using auxiliary oral hygiene devices, 38.9% of the dental hygiene students said it's bothersome, and 56.7% of the other department students said they didn't know about these things(p < 0.05). 10. In the recognition level of dental floss, in knowing the way to use it, whether or not of using it, the students of both department showed a significant difference (p < 0.05). 11. In the recognition level of floss holder and the yes or not of using it, there was an insignificant difference(p > 0.05), while in understanding of it, there was a significant difference(p < 0.05). 12. Relating to the recognition level of electric tooth brush, yes or no of using it, there wasn't a significant difference(p > 0.05). 13. In the cognition level of rubber stimulator, yes or no of using it, there wasn't much difference, while in understanding to use it, there was a significant difference(p < 0.05). 14. In the understanding level of water pik, yes or not of using it, there wasn't a great deal of difference(p < 0.05), but in inquiry of whether or not of using it, there was no difference (p > 0.05), 15. In the understanding level of tongue scraper, whether or not of using it, there wasn't much difference(p > 0.05), but in understanding to use it, there was a great deal of difference(p < 0.05). 16. In the recognition of garglin, understanding of it, and whether or not of using it, the two departments didn't show a significant difference(p < 0.05).

• Korean Journal of Remote Sensing
• /
• v.35 no.3
• /
• pp.457-470
• /
• 2019

Automatic Target Recognition (ATR) using Synthetic Aperture Radar (SAR) has been attracted attention in the fields of surveillance, reconnaissance, and national security due to its advantage of all-weather and day-and-night imaging capabilities. However, there have been some difficulties in automatically identifying targets in real situation due to various observational and environmental conditions. In this paper, ATR problems in Extended Operating Conditions (EOC) were investigated. In particular, we considered partial occlusions of the target (10% to 50%) and differences in the depression angle between training ($17^{\circ}$) and test data ($30^{\circ}$ and $45^{\circ}$). To simulate various occlusion conditions, SARBake algorithm was applied to Moving and Stationary Target Acquisition and Recognition (MSTAR) images. The ATR accuracies were evaluated by using the template matching and Adaboost algorithms. Experimental results on the depression angle showed that the target identification rate of the two algorithms decreased by more than 30% from the depression angle of $45^{\circ}$ to $30^{\circ}$. The accuracy of template matching was about 75.88% while Adaboost showed better results with an accuracy of about 86.80%. In the case of partial occlusion, the accuracy of template matching decreased significantly even in the slight occlusion (from 95.77% under no occlusion to 52.69% under 10% occlusion). The Adaboost algorithm showed better performance with an accuracy of 85.16% in no occlusion condition and 68.48% in 10% occlusion condition. Even in the 50% occlusion condition, the Adaboost provided an accuracy of 52.48%, which was much higher than the template matching (less than 30% under 50% occlusion).

• Food Science and Preservation
• /
• v.17 no.1
• /
• pp.79-83
• /
• 2010

The study was conducted to analyze the tag reading and box recognition performance of a gate-type RFID system using the UHF band to detect containers of red pepper. The reading rate of tags attached to container boxes was higher as tags were closer to antennas and the number of antennas was increased. Under optimal conditions, the reading rate was 100% and the range of distance from a carton to an antenna was 1-4 meters. When tags were attached to two sides of a box, the reading rate was lower when the tags were attached at the front and side. This was caused by data collision problems between tags. The reading rate of tags was 71.1-77.8% and the reading rate of red pepper boxes was 97.8-100.0% when the distance between the pallet under the boxes and four units of antennas was 5 meters or less, and when tags were attached at the front and side of boxes.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.37 no.3
• /
• pp.12-26
• /
• 2000

In this paper, we propose a post-processing algorithm to remove false minutiae which decrease the overall performance of an automatic fingerprint identification system by increasing computational complexity, FAR(False Acceptance Rate), and FRR(False Rejection Rate) in matching process. The proposed algorithm extracts candidate minutiae from thinned fingerprint image. Considering characteristics of the thinned fingerprint image, the algorithm selects the minutiae that may be false and located in recoverable area. If the area where the selected minutiae reside is thinned incorrectly due to noise and loss of information, the algorithm recovers the area and the selected minutiae are removed from the candidate minutiae list. By examining the ridge pattern of the block where the candidate minutiae are found, true minutiae are recovered and in contrast, false minutiae are filtered out. In an experiment, Fingerprint images from NIST special database 14 are tested and the result shows that the proposed algorithm reduces the false minutiae extraction rate remarkably and increases the overall performance of an automatic fingerprint identification system.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.34 no.2
• /
• pp.303-317
• /
• 2010

This study helps in the production of brassieres suitable for female adults by researching the actual wearing conditions and purchasing status, satisfaction rate, and preferences of female college students in their early 20's. The collected data was analyzed by a SPSSWIN 13.0 Program and the results of the research are as follows. 1. There is a noticeable difference between the breast satisfaction rate and breast size, in addition the satisfactory rate was higher in the normal breast size or a little ample size than in the very small or very big size. According to the shape, the satisfactory rate for the breast appeared to be high in the case of the recognition and evaluation of the hemisphere type than the flat, cone, pop out, and downward type. 2. For fitness, the looseness at the top of the cup (pressed between the breasts at the upper part of the front middle), pressing and looseness at the upper sides of the cup, inappropriate size of the cup, tightness of the wings, tightness at the bottom round of the breasts, the narrow width of the wire, wide space of the shoulder strings, and the sliding of the shoulder strings had problems that needed improvement. 3. The major priorities for purchasing brassieres are size, fitness, and aesthetic qualities. As a result, the size and the fitness are more important than the trend or decorations since the brassiere has the function to support the breasts that shows that hygiene and sanitation are recognized as an important standard for undergarment selection. The size and the fitness are important factors regardless of breast types in the examination of the selection standards of the brassieres for each breast type, but the cone and hemisphere types have higher preferences for design; the pop out type has more considerations for the material of the cloth. The result show that appearance is more important for smaller breasts, but the functionality of the brassier is more important than the appearance for medium and larger breast sizes.

• Nuclear Engineering and Technology
• /
• v.55 no.11
• /
• pp.3940-3955
• /
• 2023

A new method of numerical simulating prediction and decontamination effect evaluation for abrasive jet decontamination to radioactively contaminated metal is proposed. Based on the Computational Fluid Dynamics and Discrete Element Model (CFD-DEM) coupled simulation model, the motion patterns and distribution of abrasives can be predicted, and the decontamination effect can be evaluated by image processing and recognition technology. The impact of three key parameters (impact distance, inlet pressure, abrasive mass flow rate) on the decontamination effect is revealed. Moreover, here are experiments of reliability verification to decontamination effect and numerical simulation methods that has been conducted. The results show that: ⁶⁰Co and other homogeneous solid solution radioactive pollutants can be removed by abrasive jet, and the average removal rate of Co exceeds 80%. It is reliable for the proposed numerical simulation and evaluation method because of the well goodness of fit between predicted value and actual values: The predicted values and actual values of the abrasive distribution diameter are Ф57 and Ф55; the total coverage rate is 26.42% and 23.50%; the average impact velocity is 81.73 m/s and 78.00 m/s. Further analysis shows that the impact distance has a significant impact on the distribution of abrasive particles on the target surface, the coverage rate of the core area increases at first, and then decreases with the increase of the impact distance of the nozzle, which reach a maximum of 14.44% at 300 mm. It is recommended to set the impact distance around 300 mm, because at this time the core area coverage of the abrasive is the largest and the impact velocity is stable at the highest speed of 81.94 m/s. The impact of the nozzle inlet pressure on the decontamination effect mainly affects the impact kinetic energy of the abrasive and has little impact on the distribution. The greater the inlet pressure, the greater the impact kinetic energy, and the stronger the decontamination ability of the abrasive. But in return, the energy consumption is higher, too. For the decontamination of radioactively contaminated metals, it is recommended to set the inlet pressure of the nozzle at around 0.6 MPa. Because most of the Co elements can be removed under this pressure. Increasing the mass and flow of abrasives appropriately can enhance the decontamination effectiveness. The total mass of abrasives per unit decontamination area is suggested to be 50 g because the core area coverage rate of the abrasive is relatively large under this condition; and the nozzle wear extent is acceptable.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.7
• /
• pp.4508-4515
• /
• 2014

Real-time body detection has been researched actively. On the other hand, the detection rate of color distorted images is low because most existing detection methods use static skin color model. Therefore, this paper proposes a new method for detecting the skin color region using a gradient map and saturation features. The basic procedure of the proposed method sequentially consists of creating a gradient map, extracting a gradient feature of skin regions, noise removal using the saturation features of skin, creating a cluster for extraction regions, detecting skin regions using cluster information, and verifying the results. This method uses features other than the color to strengthen skin detection not affected by light, race, age, individual features, etc. The results of the detection rate showed that the proposed method is 10% or more higher than the traditional methods.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.6
• /
• pp.261-268
• /
• 2015

In spoken chatting systems, users'spoken queries are converted to text queries using automatic speech recognition (ASR) engines. If the top-1 results of the ASR engines are incorrect, these errors are propagated to the spoken chatting systems. To improve the top-1 accuracies of ASR engines, we propose a post-processing model to rearrange the top-n outputs of ASR engines using a ranking support vector machine (RankSVM). On the other hand, a number of chatting sentences are needed to train chatting systems. If new chatting sentences are not frequently added to training data, responses of the chatting systems will be old-fashioned soon. To resolve this problem, we propose a data collection model to automatically select chatting sentences from TV and movie scenarios using a support vector machine (SVM). In the experiments, the post-processing model showed a higher precision of 4.4% and a higher recall rate of 6.4% compared to the baseline model (without post-processing). Then, the data collection model showed the high precision of 98.95% and the recall rate of 57.14%.

• The Korean Journal of Emergency Medical Services
• /
• v.13 no.1
• /
• pp.97-114
• /
• 2009

Purpose : This study was performed to identify the knowledge and educational wants for first-aid and its related factors of senior students in elementary schools. Methods : From July 11-25, 2008, this study surveyed 437 elementary school students using a structured questionnaire. Result : Of the study subjects, 47.6% of the subjects had experienced first-aid conditions 25.4% had injuries, 11.0% had indigestion, 10.5% had sprains, 6.2% had acute abdominal pains, 5.9% had brash, and 5.7% had fractures. The experience rates of first-aid conditions were significantly different according to gender, mothers' educational status, means of school attendance and amount of exercise. The experience rate of first-aid education was 71.4%. There was a significant difference with grade, and school reports. They had education about cardiopulmonary resuscitation (52.5%), injury treatment and desmology(48.7%) in a good order. The score of knowledge about first-aid was $13.95{\pm}4.1$ and it was converted on the basis of 100 points, the results were below 50points, and there was a significant difference in grades. Educational wants for first-aid was 64.3%, and as a result of multi-logistic regression analysis of characteristics relating to educational wants, there was a significant difference between gender, grade, mothers' educational status, and amount of exercise. Conclusion : The rate of first-aid experience of elementary school students was high, but as their recognition and knowledge on first-aid was low, the importance of first-aid education and educational wants should be reflected for practical and organized education.