Image Binarization using Intensity Range of Grayscale Images

Kwang Baek Kim 보안공학연구지원센터 2015 International Journal of Multimedia and Ubiquitous Vol.10 No.7

The performance of binarization algorithms is determined by the selection of threshold value for binarization, and most of the previous binarization algorithms analyze the intensity distribution of the original images by using the histogram and determine the threshold value using the mean value of intensity or the intensity value corresponding to the valley of the histogram. The previous algorithms could not get the proper threshold value in the case that doesn't show the bimodal characteristic in the intensity histogram or for the case that tries to separate the feature area from the original image. Therefore, this paper proposed a novel algorithm for image binarization, which, first segments the intensity range of grayscale images to several intervals and calculates a mean value of intensity for each interval, and then repeats the interval integration until getting the final threshold value. The interval integration of two neighborhood intervals calculates the ratio of the distances between mean value and adjacent boundary value of two intervals and determine as the threshold value of the new integrated interval the intensity value that divides the distance between mean values of two intervals according to the ratio. The experiment for performance evaluation showed that the proposed algorithm generates the more effective threshold value than the previous algorithms.

Document Image Binarization by GAN with Unpaired Data Training

Quang-Vinh Dang,Guee-Sang Lee 한국콘텐츠학회(IJOC) 2020 International Journal of Contents Vol.16 No.2

Data is critical in deep learning but the scarcity of data often occurs in research, especially in the preparation of the paired training data. In this paper, document image binarization with unpaired data is studied by introducing adversarial learning, excluding the need for supervised or labeled datasets. However, the simple extension of the previous unpaired training to binarization inevitably leads to poor performance compared to paired data training. Thus, a new deep learning approach is proposed by introducing a multidiversity of higher quality generated images. In this paper, a two-stage model is proposed that comprises the generative adversarial network (GAN) followed by the U-net network. In the first stage, the GAN uses the unpaired image data to create paired image data. With the second stage, the generated paired image data are passed through the U-net network for binarization. Thus, the trained U-net becomes the binarization model during the testing. The proposed model has been evaluated over the publicly available DIBCO dataset and it outperforms other techniques on unpaired training data. The paper shows the potential of using unpaired data for binarization, for the first time in the literature, which can be further improved to replace paired data training for binarization in the future.

퍼지 이론을 이용한 영상 이진화에 관한 연구

김광백 신라대학교 자연과학연구소 2003 自然科學論文集 Vol.12 No.-

The image binarization is applied frequently as the one part of the preprocessing phase for a variety of image processing techniques such as character recognition and image analysis, etc. The setting of the threshold value influences the performance of the binarization algorithms, and most of the previous binarization algorithms analyze the intensity distribution of the original images by using the histogram and determine the threshold value using the mean value of intensity or the intensity value corresponding to the valley of the histogram. The previous algorithms could not get the proper threshold value in the case that doesn't show the bimodal characteristic in the intensity histogram or for the case that tries to separate the feature area from the original image. In this paper, I proposed a method that establish dynamically threshold using triangle type membership function about images. The proposed fuzzy binarization method establishes smallest brightness value and maximum brightness value using the control rate of brightness calculating the darkest pixel value and distance of lightest pixel price via average brightness value and applies in membership function of triangle. Through the membership is applied a -cut rule about membership function, the image is binarized. The experiment for the performance evaluation of the proposed binarization algorithm showed that the proposed algorithm generates the more effective threshold value than the previous algorithms.

Stroke Width-Based Contrast Feature for Document Image Binarization

( Le Thi Khue Van ),( Guee Sang Lee ) 한국정보처리학회 2014 Journal of information processing systems Vol.10 No.1

Automatic segmentation of foreground text from the background in degraded document images is very much essential for the smooth reading of the document content and recognition tasks by machine. In this paper, we present a novel approach to the binarization of degraded document images. The proposed method uses a new local contrast feature extracted based on the stroke width of text. First, a pre-processing method is carried out for noise removal. Text boundary detection is then performed on the image constructed from the contrast feature. Then local estimation follows to extract text from the background. Finally, a refinement procedure is applied to the binarized image as a post-processing step to improve the quality of the final results. Experiments and comparisons of extracting text from degraded handwriting and machine-printed document image against some well-known binarization algorithms demonstrate the effectiveness of the proposed method.

Algorithm to Estimate Oil Spill Area Using Digital Properties of Image

Hye-Jin Jang,남종호 한국해양공학회 2020 韓國海洋工學會誌 Vol.34 No.1

Oil spill accidents at sea result in a wide range of damages, including the destruction of ocean environments and ecosystems, as well as human illnesses by the generation of harmful gases caused by phase changes in crude oil. When an oil spill occurs, an immediate initial action should be performed to minimize the potential damage. Existing studies have attempted to identify crude oil spillage by calculating the crude oil spill range using synthetic aperture radar (SAR) satellite images. However, SAR cannot capture rapidly evolving events because of its low acquisition frequency. Herein, an algorithm for estimating an oil spill area from an image obtained using a digital camera is proposed. Noise that may occur in the image when it is captured is first eliminated by preprocessing, and then the image is analyzed. After analyzing the characteristics of the digital image, a strategy to binarize an image using the color, saturation, or lightness contained in it is adopted. It is found that the oil spill area can be readily estimated from a digital image, allowing for a faster analysis than any conventional method. The usefulness of the oil spill area measurement was confirmed by applying the developed algorithm to various oil spill images.

그림자가 있는 차량 번호판의 이진화

서병훈(Byung Hoon Seo),김병만(Byeong Man Kim),문창배(Chang Bae Moon),신윤식(Yoon Sik Shin) 한국산업정보학회 2008 한국산업정보학회논문지 Vol.13 No.4

본 논문은 이동 중의 차량으로부터 획득한 후면 번호판의 이미지를 이진화 할 경우의 문제점과 이의 해결책을 제시하였다. 후면 번호판을 이진화한 경우 태양의 고도와 차량 구조의 영향으로 그림자가 드리워진 이미지를 획득하게 되며 이 번호판의 이미지를 이진화 할 경우 문자를 인식하기에 좋지 않은 이미지를 획득하게 된다. 따라서 본 논문에서는 먼저 그림자 경계선을 파악하고 이를 이용하여 그림자가 드리워진 영역과 드리워지지 않은 영역을 구분한 후 각각의 영역을 이진화하는 방법을 제시하였다. 기존 발표되었던 이진화 방법들 중 일반 이진화, 타 논문 이진화 블록 이진화, 라벨링 응용 이진화 들과 본 논문에서 제시한 방법을 비교 실험하여 성능분석 하였고, 실험결과, 대부분의 경우에 본 논문에서 제안한 방법이 타 방법에 비해 성능이 좋음을 확인할 수 있었다. In this paper, we propose a method to solve a problem in binarizing the rear number plate image captured by a camera on a moving vehicle. An image may be shadowed by the cavernous structure of the rear side of a moving vehicle and it makes us hard to get a high quality of binary image. Therefore, we first detect a shadow edge and then divide an image into the shadow part and non-shadow part by the edge. Finally, the binary image is obtained by binarizing each part and merging them. In this paper, we do comparative work on a group of binarization methods including our method, the method suggested by Zheng, the method using block binarization, and the method using labeling. The result shows that our method achieves better performance than others in most cases.

A Review of Image Analysis in Biochemical Engineering

Sang-Kyu Jung 한국생물공학회 2019 Biotechnology and Bioprocess Engineering Vol.24 No.1

The purpose of image analysis is to extract useful information from images. Since modern image analysis allows fast, accurate, and reliable quantitative analysis, it is widely used at present in many areas of research and development. In this article, I review the image analysis methods that are commonly used in biochemical engineering, for which the subjects of the image analysis vary from molecules or cells to whole animals or biomaterials. Images captured by imaging hardware, which is not limited to digital cameras, are processed in multiple steps by applying various image processing algorithms to extract quantitative features. Image analysis has been successfully applied in diverse applications, ranging from simple densitometric evaluation to animal phenotyping and biomass analysis. Although machine learning is poised to become an increasingly common method of image analysis, traditional methods utilizing blob extraction from binarized images are likely to remain in use for the foreseeable future.

Efficient Document Image Binarization

Kiju Lee,Soowoong Jeong,Jeong-su Oh,Jong-soo Choi,Sangkeun Lee 중앙대학교 영상콘텐츠융합연구소 2014 TechArt :Journal of Arts and Imaging Science Vol.1 No.4

In this paper, we present an efficient and adaptive method for document image binarization using degraded images. Our contributions lead to improvements in both the binarization quality and speed. The vertical and horizontal strokes of Korean and Chinese characters are detected adaptively using a local means for accurate binarization. A down-sampled integral image is employed to quickly compute the local means. The proposed method can deal with various degradations owing to shadows, non-uniform illumination, low contrast, and high levels of signal-dependent noise. The experimental results show that the proposed method is about 11- to 424-times faster than existing algorithms and provides better quality results. Therefore, the proposed efficient document binarization algorithm can be a useful tool in fast documentation and optical character recognition.

Stroke Width-Based Contrast Feature for Document Image Binarization

Van, Le Thi Khue,Lee, Gueesang Korea Information Processing Society 2014 Journal of information processing systems Vol.10 No.1

Automatic segmentation of foreground text from the background in degraded document images is very much essential for the smooth reading of the document content and recognition tasks by machine. In this paper, we present a novel approach to the binarization of degraded document images. The proposed method uses a new local contrast feature extracted based on the stroke width of text. First, a pre-processing method is carried out for noise removal. Text boundary detection is then performed on the image constructed from the contrast feature. Then local estimation follows to extract text from the background. Finally, a refinement procedure is applied to the binarized image as a post-processing step to improve the quality of the final results. Experiments and comparisons of extracting text from degraded handwriting and machine-printed document image against some well-known binarization algorithms demonstrate the effectiveness of the proposed method.

A Fast Algorithm for Korean Text Extraction and Segmentation from Subway Signboard Images Utilizing Smartphone Sensors

Igor Milevskiy,Jin-Young Ha 한국정보과학회 2011 Journal of Computing Science and Engineering Vol.5 No.3

We present a fast algorithm for Korean text extraction and segmentation from subway signboards using smart phone sensors in order to minimize computational time and memory usage. The algorithm can be used as preprocessing steps for optical character recognition (OCR): binarization, text location, and segmentation. An image of a signboard captured by smart phone camera while holding smart phone by an arbitrary angle is rotated by the detected angle, as if the image was taken by holding a smart phone horizontally. Binarization is only performed once on the subset of connected components instead of the whole image area, resulting in a large reduction in computational time. Text location is guided by user’'s marker-line placed over the region of interest in binarized image via smart phone touch screen. Then, text segmentation utilizes the data of connected components received in the binarization step, and cuts the string into individual images for designated characters. The resulting data could be used as OCR input, hence solving the most difficult part of OCR on text area included in natural scene images. The experimental results showed that the binarization algorithm of our method is 3.5 and 3.7 times faster than Niblack and Sauvola adaptive-thresholding algorithms, respectively. In addition, our method achieved better quality than other methods.