• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.12
• /
• pp.81-89
• /
• 1997

In this paper, we propose a new image retrieval algorithm using the block classification. More specifically, we classify nonoverlappint small image blocks into texture, monotone, and various edges. Using these classification results and the RGB color histogram, we propose a new similarity measure which considers both local and global fretures. According to our experimental results using 232 color images, the retrieval efficiencies of the proposed and the previous methods were 0.610 and 0.522, respectively, which implies that the proposed algorithm yields better performance.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.1409-1412
• /
• 1987

Median filter techniques were applied to the chest radiographs for the digital image processing. Histogram distributions were transformed to convert the gray level for detections of the small gray deviations of the image boundaries.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.1972-1975
• /
• 2003

Video segmentation is an essential part in region-based video coding and any other fields of the video processing. Among lots of methods proposed so far, the watershed method in which the region growing is performed for the gradient image can produce well-partitioned regions globally without any influence on local noise and extracts accurate boundaries. But, it generates a great number of small regions, which we call over segmentation problem. Therefore we proposes image segmentation improvement by selective application structuring element of mathematical morphology.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.443-444
• /
• 2006

In this paper, a spatial-based perceptual watermarking considering human visual system (HVS) that is proposed for small-size images such as internet shopping-mall image. In our method, a multi-bit data can be embedded in luminance component of color images still keeping the perceptual quality of image. Experimental results demonstrated that watermarks can be strongly embedded while preserving a good fidelity.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.5
• /
• pp.103-111
• /
• 2007

In this paper, we propose a VLSI architecture of hardware optimized high quality image signal processor for a Single-chip CMOS Image Sensor(CIS). The Single-chip CIS is usually used for mobile applications, so it has to be implemented as small as possible while maintaining the image quality. Several image processing algorithms are used in ISP to improve captured image quality. Among the several image processing blocks, demosaicing and image filter are the core blocks in ISP. These blocks need line memories, but the number of line memories is limited in a low cost Single-chip CIS. In our design, high quality edge-adaptive and cross channel correlation considered demosaicing algorithm is adopted. To minimize the number of required line memories for image filter, we share the line memories using the characteristics of demosaicing algorithm which consider the cross correlation. Based on the proposed method, we can achieve both high quality and low hardware complexity with a small number of line memories. The proposed method was implemented and verified successfully using verilog HDL and FPGA. It was synthesized to gate-level circuits using 0.25um CMOS standard cell library. The total logic gate count is 37K, and seven and half line memories are used.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2006.05a
• /
• pp.45-46
• /
• 2006

In this paper, a dual stage servo mechanism has been developed for image tracking system to improve transient control performances such as small rise time, small overshoot, small settling time, etc. A secondary stage, a platform, actuated by a pair of electro-magnets is mounted on a conventional elevation gimbal. In this mechanism, the gimbal provides large range but slow motion and the platform provides small range but fast positioning. A sliding mode control is applied to the platform positioning to attain robust performances and stability in the presence of the disturbance related to dynamic coupling of the gimbal and the platform. Results from experiments illustrate that the suggested dual stage mechanism controlled by the sliding mode control is effective in improving transient responses and attenuating the disturbance related with dynamic coupling.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.1
• /
• pp.75-80
• /
• 2021

Artificial intelligence is becoming an important part of our lives providing incredible benefits. In this respect, facial expression recognition has been one of the hot topics among computer vision researchers in recent decades. Classifying small dataset of low resolution images requires the development of a new small scale deep CNN model. To do this, we propose a method suitable for small datasets. Compared to the traditional deep CNN models, this model uses only a fraction of the memory in terms of total learnable weights, but it shows very similar results for the FER2013 and FERPlus datasets.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.5
• /
• pp.265-272
• /
• 2022

As the risk of lung cancer has increased, early-stage detection and treatment of cancers have received a lot of attention. Among various medical imaging approaches, computer tomography (CT) has been widely utilized to examine the size and growth rate of lung nodules. However, the process of manual examination is a time-consuming task, and it causes physical and mental fatigue for medical professionals. Recently, many computer-aided diagnostic methods have been proposed to reduce the workload of medical professionals. In recent studies, encoder-decoder architectures have shown reliable performances in medical image segmentation, and it is adopted to predict lesion candidates. However, localizing nodules in lung CT images is a challenging problem due to the extremely small sizes and unstructured shapes of nodules. To solve these problems, we utilize atrous spatial pyramid pooling (ASPP) to minimize the loss of information for a general U-Net baseline model to extract rich representations from various receptive fields. Moreover, we propose mixed-up attention mechanism of reverse, boundary and convolutional block attention module (CBAM) to improve the accuracy of segmentation small scale of various shapes. The performance of the proposed model is compared with several previous attention mechanisms on the LIDC-IDRI dataset, and experimental results demonstrate that reverse, boundary, and CBAM (RB-CBAM) are effective in the segmentation of small nodules.

• Journal of Korea Multimedia Society
• /
• v.15 no.11
• /
• pp.1305-1318
• /
• 2012

Image segmentation is one of the difficult research problems in machine vision and pattern recognition field. A commonly used segmentation method is the Otsu method. It is simpler and easier to implement but it fails if the histogram is unimodal or similar to unimodal. And if some target area is smaller than background object, then its histogram has the distribution close to unimodal. In this paper, we proposed an improved image segmentation method based on 1D Otsu method for a small target detection. To overcome drawbacks by unimodal histogram effect, we depressed the background histogram using a logarithm function. And to improve a signal to noise ratio, we used a local average value by the neighbor window for thresholding using 1D Otsu method. The experimental results show that our proposed algorithm performs better segmentation result than a traditional 1D Otsu method, and needs much less computational time than that of the 2D Otsu method.

• Journal of the Korean Society of Radiology
• /
• v.15 no.4
• /
• pp.525-531
• /
• 2021

The detector of the positron emission tomography (PET) is composed using a plurality of scintillation pixels and photo sensors. The use of multiple photo sensors increases cost and complicates signal processing. In this study, a detector with reduced cost and simple signal processing was designed using a small number of photo sensors. A scintillation pixel and a small number of photo sensors were used, and a optical guide was used to deliver light to all the photo sensors. A reflector is applied to the scintillation pixel and the optical guide to transmit the maximum amount of light to the photo sensor. A diffuse reflector and a specular reflector were used for the reflector, and a flood image was obtained by applying different thicknesses of the optical guide. An optimal combination was selected through comparative analysis of the acquired flood images. As a result, when specular reflectors were used for both the scintillation pixel and the optical guide, excellent flood images were obtained from optical guides of all thicknesses. For the optical guide, the optimal image was obtained when using a 3 mm thickness in consideration of the size of the image and the analysis of the point where the image of the scintillation pixel was formed.