• 대한의용생체공학회:의공학회지
• /
• 제44권4호
• /
• pp.293-301
• /
• 2023

Scoliosis is a three-dimensional deformation of the spine that is a deformity induced by physical or disease-related causes as the spine is rotated abnormally. Early detection has a significant influence on the possibility of nonsurgical treatment. To train a deep learning model with preprocessed images and to evaluate the results with and without data augmentation to enable the diagnosis of scoliosis based only on a chest X-ray image. The preprocessed images in which only the spine, rib contours, and some hard tissues were left from the original chest image, were used for learning along with the original images, and three CNN(Convolutional Neural Networks) models (VGG16, ResNet152, and EfficientNet) were selected to proceed with training. The results obtained by training with the preprocessed images showed a superior accuracy to those obtained by training with the original image. When the scoliosis image was added through data augmentation, the accuracy was further improved, ultimately achieving a classification accuracy of 93.56% with the ResNet152 model using test data. Through supplementation with future research, the method proposed herein is expected to allow the early diagnosis of scoliosis as well as cost reduction by reducing the burden of additional radiographic imaging for disease detection.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.799-804
• /
• 1990

In the blade grinding process, Blade image is captured. It is captured in the environment with vibration, using monochrome CCD camera with high speed electronic shutter. The image is preprocessed using LoG filter and zero crossing. We used Hough transformation to detect straight lines from the preprocessed image. From the equations of detected lines, we calculated width between lines caused by grinders. This paper proposes automatic measurement of blade width to automate the process control of blade grinding line.

• 인지과학
• /
• 제25권1호
• /
• pp.51-71
• /
• 2014

본 논문에서는 도로 주행에서 취득한 영상을 개선하는 방법을 제안한다. 일반적인 도로주행 영상은 다양한 조명 환경과 날씨 상태로 인하여 선명하지 못한 영상이 취득되기도 한다. 특히 역광이나 야간에는 품질이 좋은 선명한 영상을 얻기가 더욱 어려우며, 이는 비전기반 지능형 자동차 기술의 응용에 많은 어려움을 준다. 인간의 시각 인지방법은 여러 가지조명 조건을 고려하여 색을 지각한다. 하지만 기존의 영상 개선 방법들은 광원의 위치와 광도, 기하학적 관계를 고려하지 않기 때문에 완벽한 결과를 얻기가 어려우며, 오히려 영상의 질이 떨어지는 경우도 발생한다. 본 논문에서는 이러한 문제점을 해결하기 위해서 1) 주어진 입력 영상의 전처리 과정을 수행한 후, 2) 선명도를 추정하여 색채의 대비를 평가하고, 3) 과대 및 과소평가 결과를 전처리된 영상과 혼합하여 사람이 지각하는 색상과 같이 개선된 영상을 얻는 효과적인 방법을 제안한다. 본 논문에서 제안하는 방법은 시각적으로 개선된 결과를 보여줄 뿐만 아니라 비전기반 지능형 자동차 기술의 한 응용분야인 교통표지판 검출의 전처리 과정으로 적용되어 성능이 향상됨을 확인할 수 있었다.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2003년도 ISIS 2003
• /
• pp.42-45
• /
• 2003

This paper describes a new iris recognition method using shift-invariant subbands. First an iris image is preprocessed to compensate the variation of the iris image. Then, the preprocessed iris image is decomposed into multiple subbands using a shift invariant wavelet transform. The best subband among them, which have rich information for various iris pattern and robust to noises, is selected for iris recognition. The quantized pixels of the best subband yield the feature representation. Experimentally, we show that the proposed method produced superb performance in iris recognition.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.2485-2489
• /
• 2005

This paper introduces a new preprocessing scheme to replace original data of gray scale images with particular ordered data so that performance of lossless compression can be improved more efficiently. As a kind of preprocessing technique to maximize performance of entropy encoder, the proposed method converts the input image data into more compressible form. Before encoding a stream of the input image, the proposed preprocessor counts co-occurrence frequencies for neighboring pixel pairs. Then, it replaces each pair of adjacent gray values with particular ordered numbers based on the investigated co-occurrence frequencies. When compressing ordered image using entropy encoder, we can expect to raise compression rate more highly because of enhanced statistical feature of the input image. In this paper, we show that lossless compression rate increased by up to 37.85% when comparing results from compressing preprocessed and non-preprocessed image data using entropy encoder such as Huffman, Arithmetic encoder.

• 제어로봇시스템학회논문지
• /
• 제8권7호
• /
• pp.589-598
• /
• 2002

The color measuring/matching expert system is implemented with a new color measuring method that combines intelligent algorithms with image processing techniques. Color measuring part of the proposed system preprocesses the scanned original color input images to eliminate their distorted components by means of the image histogram technique of image pixels, and then extracts RGB(Red, Green, Blue)data among color information from preprocessed color input images. If the extracted RGB color data does not exist on the matching recipe databases, we can measure the colors for the user who want to implement the model that can search the rules for the color mixing information, using the intelligent modeling techniques such as fuzzy inference system and adaptive neuro-fuzzy inference system. Color matching part can easily choose images close to the original color for the user by comparing information of preprocessed color real input images with data-based measuring recipe information of the expert, from the viewpoint of the delta Eformula used in practical process.

• 제어로봇시스템학회논문지
• /
• 제20권9호
• /
• pp.950-956
• /
• 2014

In this paper, the real-time detection method of a DDP (Driving Direction Point) is proposed for an unmanned vehicle to safely follow the center of the road. Since the DDP is defined as a center point between two lanes, the lane is first detected using a web camera. For robust detection of the lane, the binary thresholding and the labeling methods are applied to the color camera image as image preprocessing. From the preprocessed image, the lane is detected, taking the intrinsic characteristics of the lane such as width into consideration. If both lanes are detected, the DDP can be directly obtained from the preprocessed image. However, if one lane is detected, the DDP is obtained from the inverse perspective image to guarantee reliability. To verify the proposed method, several experiments to detect the DDPs are carried out using a 4 wheeled vehicle ERP-42 with a web camera.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 신호처리소사이어티 추계학술대회 논문집
• /
• pp.27-30
• /
• 2003

In this paper a new approach of digitalized ‘Takbon’ is introduced. By image signal processing, the letters which were written on stones can be deciphered. Epigraphic letter is detected by digital image device, digital camera. The two dimensional digital image is preprocessed because of sensor noise and detective turbulence. Color image is transformed into grey level. The letter image is analyzed in time/frequency domain. By the resultant analysis data decisive functions are calculated. Signal Processing techniques, such as scaling, clipping, digital negative, high/low filter, morphology and so on, provide algorithms that can extract letter from stones.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 하계종합학술대회 논문집(4)
• /
• pp.359-362
• /
• 2002

In this paper, we propose a new feature-based text watermarking for the binary text image. The structure of specific characters from preprocessed text image are modified to embed watermark. Watermark message are embedded and detected by the following method; Hole line disconnect using the connectivity of the character containing a hole, Center line shift using the hole area and Differential encoding using difference of flippable score points. Experimental results show that the proposed method is robust to rotation and scaling distortion.

• Journal of Information Processing Systems
• /
• 제19권4호
• /
• pp.465-473
• /
• 2023

An image enhancement algorithm and image defogging method are studied in this paper. The formation of fog and the characteristics of fog image are analyzed, and the fog image is preprocessed by histogram equalization method; then the additive white noise is removed by foggy image attenuation model, the atmospheric scattering physical model is constructed, the image detail characteristics are enhanced by image enhancement method, and the visual effect of defogging image is enhanced by guided filtering method. The proposed method has a good defogging effect on the image. When the number of training iterations is 3,000, the peak signal-to-noise ratio of the proposed method is 43.29 dB and the image structure similarity is 0.9616, indicating excellent image defogging effect.