• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4439-4455
• /
• 2021

Compared with two-stage object detection algorithms, one-stage algorithms provide a better trade-off between real-time performance and accuracy. However, these methods treat the intermediate features equally, which lacks the flexibility to emphasize meaningful information for classification and location. Besides, they ignore the interaction of contextual information from different scales, which is important for medium and small objects detection. To tackle these problems, we propose an image pyramid network based on dual attention mechanism (DAIPNet), which builds an image pyramid to enrich the spatial information while emphasizing multi-scale informative features based on dual attention mechanisms for one-stage object detection. Our framework utilizes a pre-trained backbone as standard detection network, where the designed image pyramid network (IPN) is used as auxiliary network to provide complementary information. Here, the dual attention mechanism is composed of the adaptive feature fusion module (AFFM) and the progressive attention fusion module (PAFM). AFFM is designed to automatically pay attention to the feature maps with different importance from the backbone and auxiliary network, while PAFM is utilized to adaptively learn the channel attentive information in the context transfer process. Furthermore, in the IPN, we build an image pyramid to extract scale-wise features from downsampled images of different scales, where the features are further fused at different states to enrich scale-wise information and learn more comprehensive feature representations. Experimental results are shown on MS COCO dataset. Our proposed detector with a 300 × 300 input achieves superior performance of 32.6% mAP on the MS COCO test-dev compared with state-of-the-art methods.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.1
• /
• pp.161-170
• /
• 2019

Multi-product mass production small and medium enterprise factories have a wide variety of products and a large number of products, wasting manpower and expenses for inventory management. In addition, there is no way to check the status of inventory in real time, and it is suffering economic damage due to excess inventory and shortage of stock. There are many ways to build a real-time data collection environment, but most of them are difficult to afford for small and medium-sized companies. Therefore, smart factories of small and medium enterprises are faced with difficult reality and it is hard to find appropriate countermeasures. In this paper, we implemented the contents of extension of existing inventory management method through character extraction on label with barcode and QR code, which are widely adopted as current product management technology, and evaluated the effect. Technically, through preprocessing using OpenCV for automatic recognition and classification of stock labels and barcodes, which is a method for managing input and output of existing products through computer image processing, and OCR (Optical Character Recognition) function of Google vision API. And it is designed to recognize the barcode through Zbar. We propose a method to manage inventory by real-time image recognition through Raspberry Pi without using expensive equipment.

• Korean Journal of Digital Imaging in Medicine
• /
• v.9 no.2
• /
• pp.17-21
• /
• 2007

The purpose to recognize change of average pixel value of acquisition image by control panel's density and right set up method of speed (sensitivity) and exposure dose(mAs) change that dose in purpose digital flatpanel-detector. X -ray generator DHF-158H2(Hitachi, Japan). Detector CXDI 4OG(Canon, Japan), 12 : 1 grid and exposure ray 135 kVp, 250 mA, 10 ms. focus-detector distance 180 cm and used AEC mode. DICOM reflex analysis program used image J that is digital reflex analysis program that offer in United States America National Health Center(National Institutes of Health : NlH) phantom used chest phantom(Anthromorphic : Flukebrome.medicaI USA). An experiment chest phantom that consist by formation equivalence material use because density value( -3${\sim}$+3) in X-ray control panel and seep that is speed step(slow, medium, fast) each control experimentalize. image analysis reflex neted through an experiment using image j each image compare. These was change in dose according to slow, medium, fast and density's change in an experiment result. According to detector sensitivity and density condition set, dose was relationship dissimilarity 500% from 200%. The dose came highest when is density +3 to slow. and dose more increases gray scale's extent could know that rise. Could know whether how equipment set is important through this experiment. cause of disease which change by digital radiography system forward is thought to increase more, it is considered that suitable education by this and continuous interest about equipment need absolutely.

• The Research Journal of the Costume Culture
• /
• v.12 no.1
• /
• pp.121-135
• /
• 2004

The purpose of this study is to analyze the image evaluation and the preference of tailored jacket according to the body type and the degree of satisfaction with their body of middle-aged women. The subjects used for this study were three hundred and twenty three middle-aged women ranging from 35 to 50 years old. We measured their bodies and asked them report to the questionnaire. The results are as follows. 1) Middle-aged women were classified into four typical types of body. The type 1 was tall and thin and had a front view silhouette with the form of a letter X. The type 2 was the tallest and the medium of fatness and had a front view silhouette with the form of a letter H that the lower part of the body is short. The type 3 was the medium of height and width and had a front view silhouette with the form of a letter A that the upper part of the body is short compared with their height. The type 4 were the shortest and the fattest and had a front view silhouette with the form of a letter Y that the lower and upper parts of the body is the longest. 2) Middle-aged women roughly tended not to be satisfied with part of their body. The results showed that a neck was the most satisfied body part and the weight was the least satisfied as well as overall leg shape among all 13 body parts. 3) The type 1 had the highest the degree of satisfaction on their face size, upper arm girth, waist, hip girth, weight and body shape, etc., excepting hip shape among 4 body types.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.277-279
• /
• 2003

The study utilizes remote sensing as the main monitoring means. With different spatial high-resolution, multichannel ASTER remote sensing image as the main information in Beijing city zone; with regional border and statistical data as auxiliary factor a study between the thermal space distribution character and the underground medium is analyzed based on the GIS logical algorithm and synthetic analysis technology. Results show thermal forming mechanism and the rule of distribution is mainly related to the underground medium and the change of the city distribution. Different underground medium has different degree and intensity influence on the thermal space distribution. Furthermore, urban greenbelt and water areas can reduce the thermal effect and large-scale greenbelt creates green island effect. In addition, Road net, residential area, population density, heat resources and so on have some positive effect on the thermal distribution, which increase the local temperature and intensity on the other hand. It is important to study the thermal distribution and its related factors, which contributes to the plan, construction and development of the city.

• The Research Journal of the Costume Culture
• /
• v.14 no.1
• /
• pp.93-107
• /
• 2006

This study was designed from the concept that there is a need to establish the uniform for new emerging jobs in modern society. The purpose of the study was to examine dental hygienists'wearing condition, satisfaction, and design evaluation as to provide information for the development of dental hygienists' uniform. Three hundred fifteen dental hygienists completed questionnaires, and data were analyzed by descriptive statistics, factor analysis, multiple regression, ANOVA, and Duncan's test. The findings from the study were following. First, majority of dental hygienists currently wear shirts and pants two-piece uniform in white with no print. Second, dental hygienists' overall satisfaction level for their uniform was medium, and symbolic satisfaction as well as functional satisfaction did not reach the medium level while aesthetic satisfaction was above medium. The satisfaction levels were significant different between groups according to the education, clinic size, and work experience. Third, aesthetic satisfaction levels influenced on the job image as professional. Finally, designs consisting of two-piece style with shirts and pants and having Peter pan collar were recorded high evaluation. The findings can be implemented into dental hygienists' uniform design.

• Journal of Satellite, Information and Communications
• /
• v.10 no.4
• /
• pp.117-127
• /
• 2015

Recently, research and development on satellite payloads are being developed such as the optical sensor, SAR etc. Satellite image for earth observation is being utilized both domestically and abroad. Advanced satellite payload technology has led to the collection and analysis of satellite images relying on the optical sensor. Currently, related organizations such as RDA(the Rural Development Administration) are collectively collaborating to plan a national project to develop a medium-sized satellite based on Korea's domestic technology independently. This paper investigated the cases of the past research on application of satellite images for agriculture and analyzed the technical specifications for satellite payload in each area of such application. Based on the results of the past surveys and consultation studies among local experts in satellite image application, we analyzed the current trends, plans and applications of domestic and overseas R&D in satellite payloads for earth observation in agriculture, and proposed the appropriate technical specifications for developing a future medium-sized satellite for agriculture. The proposed specifications were then incorporated into a simulated satellite to examine its performance to observe the Korean farming areas. The authors anticipate that the findings of this paper will form a useful technical basis for providing the appropriate specifications for developing future medium-sized satellite payloads to be used in agriculture and forestry, and enabling the end users to efficiently utilize the satellite.

• Korean Journal of Human Ecology
• /
• v.13 no.5
• /
• pp.787-798
• /
• 2004

This research was designed to compare clothing fabric images between on- and off-line by fabric types and its characteristics. 125 subjects who have a fashion design major evaluated the fabric image of various 16 kinds of specimens on- and off-line. Frequency, x2-test, t-test, and Pearson correlation were used for statistical analysis by SPSS WIN 11.0. The results of this study are as follows: 1. The advantage of on-line shopping includes low price and purchasing convenience, etc. On-line shoppers mostly purchase clothing items, such as T-shirt and slacks. In most cases, texture of fabrics provides a main cause for the shoppers' discontentment with internet shopping. 2. The results of t-test for the difference of the evaluation score show that plain or pile weave and highly thick or thin fabrics-for example, Organdy, Corduroy, Nylon Taffeta, Plush, etc.-have a large difference between on-and off-line image. On the other hand, medium-thick twill weave-for example, Tweed, Flannel, etc.-or patterned weave-Chiffon, Dobby fabric, etc.-shows a small difference. 3. The results of correlation of the evaluation score indicate that wool twill fabrics like Tweed, Saxony, and Polafleece show a high correlation between two kinds of evaluation score. In texture preference, no correlation exists between on-line and off-line. 4. With an analysis on fabric image evaluation by fabric characteristics, smooth, shiny pink Satin was found the most positively-evaluated item in all evaluation fields except in pattern preference and individuality. On the other hand, thick olive green Corduroy was evaluated most negatively in the fields of elegance, luxury, and feminine. 5. When compared with real ones, thin fabrics provide a different on-line fabric image. For example, Nylon Taffeta and Organdy were evaluated positively on-line because of luster and pastel tone color. However, lusterless uneven cotton Seersucker was evaluated negatively. 6. Real fabrics preference is more negative than on-line ones. In addition, patterned or unique fabrics display a high agreement between the two kinds of images.

• Korean Journal of Radiology
• /
• v.22 no.1
• /
• pp.131-138
• /
• 2021

Objective: Iterative reconstruction degrades image quality. Thus, further advances in image reconstruction are necessary to overcome some limitations of this technique in low-dose computed tomography (LDCT) scan of the chest. Deep-learning image reconstruction (DLIR) is a new method used to reduce dose while maintaining image quality. The purposes of this study was to evaluate image quality and noise of LDCT scan images reconstructed with DLIR and compare with those of images reconstructed with the adaptive statistical iterative reconstruction-Veo at a level of 30% (ASiR-V 30%). Materials and Methods: This retrospective study included 58 patients who underwent LDCT scan for lung cancer screening. Datasets were reconstructed with ASiR-V 30% and DLIR at medium and high levels (DLIR-M and DLIR-H, respectively). The objective image signal and noise, which represented mean attenuation value and standard deviation in Hounsfield units for the lungs, mediastinum, liver, and background air, and subjective image contrast, image noise, and conspicuity of structures were evaluated. The differences between CT scan images subjected to ASiR-V 30%, DLIR-M, and DLIR-H were evaluated. Results: Based on the objective analysis, the image signals did not significantly differ among ASiR-V 30%, DLIR-M, and DLIR-H (p = 0.949, 0.737, 0.366, and 0.358 in the lungs, mediastinum, liver, and background air, respectively). However, the noise was significantly lower in DLIR-M and DLIR-H than in ASiR-V 30% (all p < 0.001). DLIR had higher signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) than ASiR-V 30% (p = 0.027, < 0.001, and < 0.001 in the SNR of the lungs, mediastinum, and liver, respectively; all p < 0.001 in the CNR). According to the subjective analysis, DLIR had higher image contrast and lower image noise than ASiR-V 30% (all p < 0.001). DLIR was superior to ASiR-V 30% in identifying the pulmonary arteries and veins, trachea and bronchi, lymph nodes, and pleura and pericardium (all p < 0.001). Conclusion: DLIR significantly reduced the image noise in chest LDCT scan images compared with ASiR-V 30% while maintaining superior image quality.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.11 no.5
• /
• pp.507-513
• /
• 2018

Steganography is a technique that uses hidden messages to prevent anyone apart from knowing the existence of a secret message, except the sender and trusted recipients. This paper applies 24 bit color image as cover medium. And a 24-bit color image has three components corresponding to red, green and blue. This paper proposes an image steganography method that uses Triple-A algorithm to hide the secret (Hangul) message by arbitrarily selecting the number of LSB bits and the color channel to be used. This paper divides the secret character into the chosung, jungsung and jongsung, and applies crossover, encryption and arbitrary insertion positions to enhance robustness and confidentiality. Experimental results of the proposed method show that insertion capacity and correlation are excellent and acceptable image quality level. Also, considering the image quality, it was confirmed that the size of LSB should be less than 2.