• Journal of KIISE:Software and Applications
• /
• v.34 no.3
• /
• pp.266-274
• /
• 2007

This paper proposes a new reconstruction method of high-resolution facial image from a low-resolution facial image based on top-down machine learning and recursive error back-projection. A face is represented by a linear combination of prototypes of shape and that of texture. With the shape and texture information of each pixel in a given low-resolution facial image, we can estimate optimal coefficients for a linear combination of prototypes of shape and those that of texture by solving least square minimizations. Then high-resolution facial image can be obtained by using the optimal coefficients for linear combination of the high-resolution prototypes. In addition, a recursive error back-projection procedure is applied to improve the reconstruction accuracy of high-resolution facial image. The encouraging results of the proposed method show that our method can be used to improve the performance of the face recognition by applying our method to reconstruct high-resolution facial images from low-resolution images captured at a distance.

• Korean Journal of Optics and Photonics
• /
• v.19 no.4
• /
• pp.249-254
• /
• 2008

We developed a new x-ray image sensor utilizing a reflection-mode liquid crystal panel as its sensitive element, and tested its functionality by using it to obtain an x-ray image of a printed circuit board. In the liquid crystal x-ray image sensors hitherto reported, the liquid crystal layer is in direct contact with the photoconductive film which is deposited on a glass substrate. In the fabrication of the new x-ray image sensor, a liquid crystal panel is fabricated in the first step by using a pair of glass plates of a few centimeters thicknrss. Then one of the glass substrates is ground until its thickness is reduced to about $60\;{\mu}m$. After polishing the glass plate, dielectric films for high reflectance at 630 nm, a film of amorphous selenium for photoconduction, and a transparent conductive film for electrode are deposited in sequence. The new x-ray image sensor has several merits: primarily, fabrication of a large area sensor is more easily compared with the old fashioned x-ray image sensors. Since the reflection type liquid crystal panel has a very steep response curve, the new x-ray sensor has much more sensitivity to x-rays compared with the conventional x-ray area sensor, and the radiation dosage can be reduced down to less then 20%. By combining the new x-ray sensor with CCD camera technology, real-time x-ray images can be easily captured. We report the structure, fabrication process and characteristics of the new x-ray image sensor.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.2
• /
• pp.19-25
• /
• 2023

In this paper, we propose a method for improving raw images captured in a low light condition based on deep learning considering the image signal processing. In the case of a smart phone camera, compared to a DSLR camera, the size of a lens or sensor is limited, so the noise increases and the reduces the quality of images in low light conditions. Existing deep learning-based low-light image processing methods create unnatural images in some cases since they do not consider the lens shading effect and white balance, which are major factors in the image signal processing. In this paper, pixel distances from the image center and channel average values are used to consider the lens shading effect and white balance with a deep learning model. Experiments with low-light images taken with a smart phone demonstrate that the proposed method achieves a higher peak signal to noise ratio and structural similarity index measure than the existing method by creating high-quality low-light images.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.2
• /
• pp.65-73
• /
• 2008

The characterization process for the accurate color reproduction in mobile phone with camera and LCD is popular. The camera and LCD characterization, gamut mapping process is necessary to map the camera's input color stimulus, CIEXYZ value, into the LCD's output color stimulus. Each characterization is the process estimating the relation between input and output signals. In case of LCD, because of output device, the output color stimulus for the arbitrary input signal can be measured by spectro-radiometer However, in the camera, as the input device, the characterization is an inaccurate and needs the manual works in the process obtaining the output signal because the input signal can not be generated. Moreover, after gamut mapping process, the noise is increased because the optimized gamma tone curve of camera for the noise is distorted by the characterization. Thus, this paper proposed the system of obtaining the output signal of camera and the method of gamma correction for the noise. The camera's output signal is obtained by RGB values of patches from captured the color chart image. However, besides the illumination, the error for the location of the chart in the viewfinder is generated when many camera modules are captured the chart. The method of correcting the position to correct the error from manual works. The position of camera is estimated by captured image. This process and moving of camera is accomplished repeatedly, and the optimized position can be obtained. Moreover, the lightness curve of camera output is corrected partly to reduce the noise from the characterization process.

• Physical Therapy Korea
• /
• v.18 no.4
• /
• pp.60-67
• /
• 2011

Ultrasonography (US) is a recent technique that has proven to be useful for assessing muscle thickness and guiding the rehabilitation decision-making of clinicians and researchers. The purpose of this study was to determine the inter-rater reliability of the US measurement of transversus abdominis (TrA), internal oblique (IO), and external oblique (EO) thicknesses for different probe locations and measurement techniques. Twenty healthy volunteers were recruited in this study. Muscle thicknesses of the transversus TrA, IO, and EO were measured three times in the hook-lying position. The three different probe locations were as follows: 1) Probe location 1 (PL1) was below the rib cage in direct vertical alignment with the anterior superior iliac spine (ASIS). 2) Probe location 2 (PL2) was halfway between the ASIS and the ribcage along the mid-axillary line. 3) Probe location 3 (PL3) was halfway between the iliac crest and the inferior angle of the rib cage, with adjustment to ensure the medial edge of the TrA. The two different techniques of thickness measurement from the captured images were as follows: 1) Muscle thickness was measured in the middle of the muscle belly, which was centered within the captured image (technique A; TA). 2) Muscle thickness was measured along a horizontal reference line located 2 cm apart from the medial edge of the TrA in the captured image (technique B; TB). The intraclass correlation coefficient (ICC [3,k]) was used to calculate the inter-rater reliability of the thickness measurement of TrA, IO and EO using the values from both the first and second examiner. In all three muscles, moderate to excellent reliability was found for all conditions (probe locations and measurement techniques) (ICC=.70~.97). In the PL1-TA condition, inter-rater reliability in the three muscle thicknesses was good to excellent (ICC=.85~.96). The reliability of all measurement conditions was excellent in IO (ICC=.95~.97). Therefore, the findings of this study suggest that TA can be applied to PL1 by clinicians and researchers in order to measure the thickness of abdominal muscles.

• Journal of the Korean Institute of Telematics and Electronics T
• /
• v.36T no.3
• /
• pp.26-31
• /
• 1999

In this paper, we propose a facial characteristic detection algorithm based on knowledge and object segmentation method for image communication. In this algorithm, under the condition of the same lumination and distance from the fixed video camera to human face, we capture input images of 256 $\times$ 256 of gray scale 256 level and then remove the noise using the Gaussian filter. Two images are captured with a video camera, One contains the human face; the other contains only background region without including a face. And then we get a differential image between two images. After removing noise of the differential image by eroding End dilating, divide background image into a facial image. We separate eyes, ears, a nose and a mouth after searching the edge component in the facial image. From simulation results, we have verified the efficiency of the Proposed algorithm.

• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.1
• /
• pp.95-100
• /
• 2007

In the steel & iron processing line, some characters are marked for the material management in the surface of material. This paper describes about the developed results of vision system for the recognition of material management characters, which was included in the slab image. Our vision system for the character recognition includes that CCD camera system which acquire slab image, optical transmission system which transmit captured image to the long distance, input and output system for the interface with existing system and monitoring system for the checking of recognition results. We have installed our vision system at the continuous casting line and tested. Also, we have performed inspection of durability, reliability and recognition rate. Through the testing, we have confirmed that our system have high recognition rate, 97.4%.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.6
• /
• pp.645-651
• /
• 2012

In this paper, we propose a new method for stabilizing the image captured from a camera mounted on a buoy robot. In this study, in order to solve the problem of cumulative errors and noise produced by a general gyro sensor measuring the orientation angle of the buoy robot, we propose new method for stabilizing the image. In this method, image processing techniques are combined with a newly designed target mounting mechanism that adapts to wave fluctuations. New target extraction and angle estimation techniques are introduced, along with the new mounting mechanism used for the camera and the target, which produce a stabilized image even if the buoy robot is on fluctuating waves.

• Journal of Biomedical Engineering Research
• /
• v.31 no.2
• /
• pp.114-122
• /
• 2010

This paper presents a method that registers a lymphoscintigraphy to the real image captured by a CMOS camera, which helps surgeons to easily and precisely detect sentinel lymph nodes for sentinel lymph node biopsy in breast cancer patients. The proposed method consists of two steps: pre-matching and image registration. In the first step, we localize fiducial markers in a lymphoscintigraphy and a real image of a four quadrant bar phantom by using image processing techniques, and then determines perspective transformation parameters by matching with the corresponding marker points. In the second step, we register a lymphoscintigraphy to a real images of patients by using the perspective transformation of pre-matching. To examine the accuracy of the proposed method, we conducted an experiment with a chest mock-up with radioactive markers. As a result, the euclidean distance between corresponding markers was less than 3mm. In conclusion, the present method can be used to accurately align lymphoscintigraphy and real images of patients without attached markers to patients, and then provide useful anatomical information on sentinel lymph node biopsy.

• The KIPS Transactions:PartB
• /
• v.10B no.7
• /
• pp.803-810
• /
• 2003

In this paper, we describe a fast image mosaic method for constructing a large-scale image with video image captured from cameras that are arranged in radial shape. In the first step, we adopt the phase correlation algorithm to estimate the horizontal and vertical displacement between two adjacent images. Secondly, we calculate the accurate transform matrix among those cameras with Levenberg-Marquardt method. In the last step, those images are stitched into one large scale image in real-time by applying the transform matrix to the texture mapping function of DirectX. The feature of the method is that we do not need to use special hardware devices or write machine-level programs for Implementing a real-time mosaic system since we use conventional graphic APIs (Application Programming Interfaces), DirectX for image synthesis process.