• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권2호
• /
• pp.76-85
• /
• 2004

In this paper, the contrast enhancement method of thermal image is proposed and it is the plateau equalization algorithm using local histogram for the real time display of infrared imagery. Through hardware implementing, its practicality and adequacy are proved. Examinations are executed to verify the effect of contrast enhancement by bright control and contrast control automatic to the plateau value in the manual mode, and that verified the effect of contrast enhancement in the automatic mode and the practicality in the real system. According to the experiment results, the proposed "the application of local histogram and plateau equalization algorithm for contrast enhancement of real time thermal image"in this dissertation is the verified method for the thermal imaging contrast enhancement.

• 대한음성학회지:말소리
• /
• 제60호
• /
• pp.145-164
• /
• 2006

This paper proposes class histogram equalization (CHEQ) to compensate noisy acoustic features for robust speech recognition. CHEQ aims to compensate for the acoustic mismatch between training and test speech recognition environments as well as to reduce the limitations of the conventional histogram equalization (HEQ). In contrast to HEQ, CHEQ adopts multiple class-specific distribution functions for training and test environments and equalizes the features by using their class-specific training and test distributions. According to the class-information extraction methods, CHEQ is further classified into two forms such as hard-CHEQ based on vector quantization and soft-CHEQ using the Gaussian mixture model. Experiments on the Aurora 2 database confirmed the effectiveness of CHEQ by producing a relative word error reduction of 61.17% over the baseline met-cepstral features and that of 19.62% over the conventional HEQ.

• 한국멀티미디어학회논문지
• /
• 제22권9호
• /
• pp.983-991
• /
• 2019

Histogram equalization method is a popular contrast enhancement technique. However, there are some drawbacks, namely, over-enhancement, under-enhancement, structure information loss, and noise amplification. In this paper, we propose an edge-enhancing histogram equalization algorithm while suppressing over-enhancement simultaneously. Firstly, over-enhancement is suppressed by clipping a transfer function, then, edge enhancement is achieved by using guided image filter. Experiments are carried out to evaluate the performance of the various HE algorithms. As a result, both qualitative and quantitative assessment showed that the proposed algorithm successfully suppressed over-enhancement while enhancing edges.

• 융합신호처리학회논문지
• /
• 제10권4호
• /
• pp.239-244
• /
• 2009

영상 대비 향상은 영상 처리 분야에서 중요한 역할을 한다. 히스토그램 스트레칭이나 히스토그램 균등화 등 기존 대비 향상 기법들과 히스토그램 균등화 기반의 수많은 방법들은 저대비에 소수의 화소들이 넓게 퍼져 있는 영상에 대해서 만족할만한 결과를 내지 못한다. 따라서 본 논문은 군집화 방법에 기반한 새로운 영상 대비 향상 기법을 제안한다. 히스토그램의 군집수는 원영상의 히스토그램을 분석하여 얻을 수 있다. 히스토그램 성분들을 K-means 알고리즘을 이용하여 군집화한다. 그리고 히스토그램 군집 범위와 군집의 화소수 비율을 비교하여 히스토그램 스트레칭과 히스토그램 균등화를 선택적으로 적용한다. 실험 결과로부터 제안한 방법이 기존의 대비 향상 기법들보다 더 효과적임을 확인할 수 있었다.

• 대한방사선기술학회지:방사선기술과학
• /
• 제47권1호
• /
• pp.29-37
• /
• 2024

This study aims to evaluate the performance of the U-Net based learning model that may vary depending on the histogram equalization algorithm. The subject of the experiment were 17 radiology students of this college, and 1,727 data sets in which the region of interest was set in the thyroid after acquiring ultrasound image data were used. The training set consisted of 1,383 images, the validation set consisted of 172 and the test data set consisted of 172. The equalization algorithm was divided into Histogram Equalization(HE) and Contrast Limited Adaptive Histogram Equalization(CLAHE), and according to the clip limit, it was divided into CLAHE8-1, CLAHE8-2. CLAHE8-3. Deep Learning was learned through size control, histogram equalization, Z-score normalization, and data augmentation. As a result of the experiment, the Attention U-Net showed the highest performance from CLAHE8-2 to 0.8355, and the U-Net and BSU-Net showed the highest performance from CLAHE8-3 to 0.8303 and 0.8277. In the case of mIoU, the Attention U-Net was 0.7175 in CLAHE8-2, the U-Net was 0.7098 and the BSU-Net was 0.7060 in CLAHE8-3. This study attempted to confirm the effects of U-Net, Attention U-Net, and BSU-Net models when histogram equalization is performed on ultrasound images. The increase in Clip Limit can be expected to increase the ROI match with the prediction mask by clarifying the boundaries, which affects the improvement of the contrast of the thyroid area in deep learning model learning, and consequently affects the performance improvement.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.192-194
• /
• 2003

Since the details in quasi-homogeneous region will be destroyed from the conventional global image enhancement method such as histogram equalization. This defect is caused by the saturation of gray level in equalization process. So the local histogram equalization for each quasi-homogeneous region will be used in order to improve the details in the region itself. To obtain the quasi- homogeneous regions, the original image must be segmented. Here we applied the watershed transform to the interesting image. Since the watershed transform is based on mathematical morphology, therefore, the regions touch can be effectively separated. Hence two adjacent regions which have the similar gray pixels will be split off. The process will be independently applied to three different spectral images. Then three different colors are assigned to each processed image in order to produce a color composite image. By the proposed algorithm, the result image shows the better perception on image details. Therefore, the high efficiency of image classification can be obtained by using this color image.

• International Journal of Advanced Culture Technology
• /
• 제3권2호
• /
• pp.161-170
• /
• 2015

Mean separated histogram equalization in order to preserve the original mean brightness has been proposed. To provide the minimum mean brightness error after the histogram modification, the input image's histogram is successively divided by the factor of 2 until the mean brightness error is satisfied the defined threshold. Then each divided group or sub-histogram will be independently equalized based on the proportional input mean. To provide the overall minimum mean brightness error, each group will be controlled by adding some certain pixels from the adjacent grey level of the next group for giving its mean near by the corresponding the divided mean. However, it still exists some little error which will be put into the next adjacent group. By successive dividing the original histogram, we found that the absolute mean brightness error is gradually decreased when the number of group is increased. Therefore, the error threshold is assigned in order to automatically dividing the original histogram for obtaining the desired absolute mean brightness error (AMBE). This process will be applied to the color image by treating each color independently.

• 한국지능시스템학회논문지
• /
• 제24권5호
• /
• pp.504-511
• /
• 2014

본 논문에서는 영상의 비선형 평활화와 다차원의 명암변화에 기반을 둔 조합형 인식기법을 제안하였다. 여기서 비선형 평활화는 적응적 변형의 히스토그램 재조정 전처리 기법으로 영상의 밝기를 조정하여 화질을 개선하기 위함이다. 다차원의 명암변화는 인접 픽셀간의 밝기변화를 4단계로 나누어 고려함으로써 영상의 속성을 더욱 더 정확하게 반영하기 위함이고, x축과 y축의 2방향 각각의 명암변화를 고려한 정규상호상관계수는 좀 더 포괄적으로 영상의 유사성을 측정하기 위함이다. 제안된 기법을 50개 40*40 픽셀의 명암도 변화를 가지는 얼굴영상들을 대상으로 실험한 결과, 평활화를 수행하지 않거나 선형 평활화를 수행한 기법에 비해 각각 영상의 속성을 잘 반영한 우수한 인식성능이 있음을 확인하였다.

• 한국지능시스템학회논문지
• /
• 제20권1호
• /
• pp.30-35
• /
• 2010

본 논문에서는 대칭 로지스틱 함수에 기반을 둔 히스토그램 평활화를 이용한 영상의 화질개선을 제안하였다. 여기서 히스토그램 평활화는 영상의 명암도를 조정하여 화질을 개선하는 간단하고 효과적인 공간영역 기반 처리기법이다. 또한 대칭 로지스틱 함수는 s-자 형의 비선형 변환함수로 영상의 명암도 발생빈도수에 따라 밝기개선 정도를 비선형적으로 조정하기 위함이다. 특히 영상의 히스토그램에서 최대 발생빈도수를 가지는 명암도와 전체 픽셀수만을 이용한 유연한 대칭의 로지스틱 함수를 제안함으로써, 기존 로지스틱 함수에서의 지수함수 계산 부담을 감소시켰다. 제안된 평활화 기법을 크기와 히스토그램 분포가 다른 5개의 영상을 대상으로 실험한 결과, 원 영상이나 기존의 전역 히스토그램 평활화의 결과영상보다 우수한 화질개선 성능이 있음을 확인하였다.

• 대한방사선기술학회지:방사선기술과학
• /
• 제44권5호
• /
• pp.473-480
• /
• 2021

This study was purpose to method that applies for improving the image quality in CT and X-ray scan, especially in the lung region. Also, we researched the parameters of the image before and after applying for Histogram Equalization (HE) such as mean, median values in the histogram. These techniques are mainly used for all type of medical images such as for Chest X-ray, Low-Dose Computed Tomography (CT). These are also used to intensify tiny anatomies like vessels, lung nodules, airways and pulmonary fissures. The proposed techniques consist of two main steps using the MATLAB software (R2021a). First, the technique should apply for the process of normalization for improving the basic image more correctly. In the next, the technique actively rearranges the intensity of the image contrast. Second, the Contrast Limited Adaptive Histogram Equalization (CLAHE) method was used for enhancing small details, textures and local contrast of the image. As a result, this paper shows the modern and improved techniques of HE and some advantages of the technique on the traditional HE. Therefore, this paper concludes that various techniques related to the HE can be helpful for many processes, especially image pre-processing for Machine Learning (ML), Deep Learning (DL).