Morphological Closing Operations Research Articles

Open-pit mine change detection from high resolution remote sensing images using DA-UNet++ and object-based approach

ABSTRACT Automatic extraction of land use and land cover changes in mining areas is crucial for the supervision of mineral resources. This study proposes an automatic open-pit mine (OM) change detection method via a novel deep learning model, DA-UNet++ (Deformable-Attention-UNet++), and an object-based approach from high resolution remote sensing images. First, by embedding a deformable convolution module and attention mechanism module into UNet++ network, and constructing a skip sampling strategy, DA-UNet++ is proposed to combat the high heterogeneity of shape and scale of ground objects in OM environment and improve the change detection accuracy. Second, an object-based post-processing (OBPP) approach is presented for enhancing raw change detection results of DA-UNet++, which includes object-based expectation maximisation optimisation and morphological close operation. In the experiments, Season-varying Change Detection Dataset is first employed to evaluate the effectiveness of the proposed DA-UNet++ by comparing with other benchmark models. Then, self-made Ordos OM Change Detection Dataset including 38 OMs is utilised to verify the efficacy of the proposed method, and comprehensive metric of 85.6% is achieved. Evaluation results prove that the proposed method provides superior change detection performance compared with several benchmark methods, and has practical significance for OM management and monitoring.

A novel algorithm for detection of tuberculosis bacilli in sputum smear fluorescence images

This work proposes an algorithm aimed at recognizing and accounting Koch bacilli in digital images of microbiological sputum samples stained with auramine, in order to determine the degree of concentration and the state of the disease (tuberculosis). The algorithm was developed with the main objective of maximizing the sensitivity and specificity of the analysis of microbiological samples (recognition and counting of bacilli) according to each preparation method (direct and diluted pellets) in order to reduce the subjectivity of the visual inspection applied by the specialist at the time of analyzing the samples. The proposed algorithm consists of a background removal, an image improvement stage based on consecutive morphological closing operations, a segmentation stage of objects of interest based on thresholdization and a classification stage based on SVM. Each algorithmic stage was developed taking into account the method of preparation of the sample to be processed, being this aspect the main contribution of the proposed work, since it was possible to achieve very satisfactory results in terms of specificity and sensitivity. In this context, sensitivity levels of 91.24% and 93.79% were obtained. Specificity levels of 90.33% and 94.85% were also achieved for direct and diluted pellet methods respectively.

A Computationally Efficient Skull Scraping Approach for Brain MR Image

Background: In the process of volumetric evaluation of the damaged region in the human brain from a MR image it is very crucial to remove the non-brain tissue from the acquainted image. At times there is a chance during the process of assessing the damaged region through automated approaches might misinterpret the non-brain tissues like skull as damaged region due to their similar intensity features. So in order to address such issues all such artefacts. Objective: In order to mechanize an efficient approach that can effectively address the issue of removing the non-brain tissues with minimal computation effort and precise accuracy. It is very essential to keep the computational time to be as minimal as possible because the processes of skull removal is used in conjunction with segmentation algorithm, and if the skull scrapping approach has consumed a considerable amount of time, they it would impact the over segmentation and volume assessment time which is not advisable. Methods: In this paper a completely novel approach named Structural Augmentation has been proposed, that could efficiently remove the skull region from the MR image. The proposed approach has several phases that include applying of Hybridized Contra harmonic and Otsu AWBF filtering for noise removal and threshold approximation through Otsu based approach and constructing the bit map based on the approximated threshold. Morphological close operation followed by morphological open operation with reference to a structural element through the generated bitmap image. Results: The experiment are carry forwarded on a real time MR images of the patient at KGH hospital, Visakhapatnam and the images from open sources repositories like fmri. The experiment is conducted on the images of varied noise variance that are tabulated in the results and implementation section of the article. The accuracy of the proposed method has been evaluated through metrics like Accuracy, Sensitivity, Specificity through true positive, true negative, False Positive and False negative evaluations. And it is observed that the performance of the proposed algorithm seems to be reasonable good. Conclusion: The skull scrapping through structural Augmentation is computationally efficient when compared with other conventional approaches concerning both computational complexity and the accuracy that could be observed on experimentation. The Adaptive Weighted Bilateral Filter that acquire the weight value from the approximated contra harmonic mean will assist in efficient removal of poison noised by preserving the edge information and Otsu algorithm is used to determine the appropriate threshold value for constructing the bitmap image of the original MRI image which is efficient over the earlier mean based approach for estimating the threshold. Moreover, the efficiency of the proposed approach could be further improved by using customized structural elements and incorporating the fuzzy based assignments among the pixels that belong to brain tissue and skull effectively.

Intelligent Bimodal Segmentation Applied on CFA Images of Food Products

Images segmentation is an essential step in the non-destructive analysis of food products by image analysis. It participates upstream in the operations of classification, identification or extraction of image attributes. To implement our segmentation approach, we extracted from each CFA image
 The R, G, B (for the RGB color space), and H, S, V (for the HSV color space). We then calculate the histogram of each of the color components. The component S for which the bimodal histogram was the most pronounced was selected and the binary mask was created from this component. The created binary mask has been filtered using morphological closing and opening operations. The application of each binary mask to its image CFA host, allowed us to separate with success the pixels representing the food products from the pixels of the background of the image CFA.

Diabetic Retinopathy Identification using Deep Believe Network

Diabetic retinopathy is an eye disease caused by swelling in blood vessels. If it gets worse the blood vessels will rupture and this is the main factor of blindness. This disease is rarely found in children and symptoms can be seen during puberty. One way to find out this disease is a wide and comprehensive pupil jerky eye examination performed by an ophthalmologist manually. Manual examination takes a long time and can cause misidentification because the similarity of characteristics of diabetic retinopathy is difficult to see directly. Because of this problem a method is needed that can facilitate the eye doctor in identifying diabetic retinopathy by simply inserting a retinal image into the system. The method proposed in this study is Deep Belief Network (DBN). The steps taken before identification are the image of the retina undergoing pre-processing in the form of grayscale, median filter, contrast stretching, morphological close operation and feature extraction using the Grey level Counselling Matrix (GLCM). In this study it was shown that the proposed method was able to identify diabetic retinopathy with 84% accuracy, sensitivity 93%, and specificity 70%.

Identification of Retinoblastoma Using Backpropagation Neural Network

Retinoblastoma (eye cancer) is an eye disease that is usually suffered by children that attack the thin nerve tissue behind the eyes (the part which is sensitive to light). Retinoblastoma can attack one or both eyes and it is a type of disease that can be caused by a genetic mutation called Retinoblastoma1 (RB1). On manual physical examination using ophthalmoscopy by a doctor or an expert there is a yellowish white / white cancer on the fundus that is often caused by the vascularization. That is why it needs a method that can be done to identify retinoblastoma disease through retinal fundus images automatically. In this research the method used is Backpropagation Neural Network using input of retinal fundus image. The stages which is done to identify retinoblastoma disease are image processing (resize, grey scaling, morphological close operation, and optic disk elimination), feature extraction using Grey Level Co-occurrence Matrix method and then classification using backpropagation neural network. After testing on the system in this research, it was concluded that the method used is able to identify retinoblastoma disease with accuracy 90%.

Inland water body extraction in complex reliefs from Sentinel-1 satellite data

Water body classification is a topic of great interest, especially for the effective management of floods. Synthetic aperture radar (SAR) imaging has demonstrated a great potential for water monitoring, given its capacity to register images independent of weather conditions. Several algorithms for water detection using SAR images are based on optimal thresholding techniques. However, these simple methodologies produce false classification results when small water bodies embedded in mountain ranges are presented in the image. We present an unsupervised and easy-to-implement methodology, based on local Moran index of spatial association in combination with morphological closing operations, for inland water body extraction. According to several experiments, we demonstrate that our method is capable of effectively extracting lakes and rivers located at different land surface reliefs without the requirement of a training step. In addition, comparisons with the state-of-the-art techniques demonstrate the effectiveness of our procedure, performing an overall accuracy of 96.37% and Kappa = 0.927.

Copy-move forgery detection using binary discriminant features

Digital images have numerous applications in various fields such as cyber forensics, courts, newspapers, magazines, medical field and many more. Therefore, the trustfulness of the contents of an image is very essential. Image manipulation has become very much simple and common due to the development of high resolution digital cameras, powerful computers and numerous image editing software. Hence, falsification of visual content of images is not at all limited to specialists and detection of forged images have become a highly challenging task. Of the different kinds of image forgery, the proposed method deals with one particular kind known as copy-move forgery. The method is an integration of the conventional block-based and keypoint based methods. The method first identifies the forged keypoint locations and then localizes the forged region using a region extraction technique. To improve the accuracy of detection, a Binary Discriminative Feature descriptor (BDF) is used for feature detection and matching. The suspected regions are identified by replacing the matching feature points with corresponding superpixel blocks. The neighbouring blocks are then merged with the suspected regions based on color similarity, using Color Histogram Matching and a final morphological close operation extracts the detected forged regions. Experimental results show that the proposed method has better detection accuracy, in terms of precision, recall and F1 Score, when compared with the state-of-the-art methods of copy-move forgery detection, under conditions of plain copy-move forgery as well as post-processing conditions like brightness changes, contrast adjustments, color reduction and blurring.

Metal stamping character recognition algorithm based on multi-directional illumination image fusion enhancement technology

Metal stamping character (MSC) automatic identification technology plays an important role in industrial automation. To improve the accuracy and stability of segmentation and recognition for MSCs, an algorithm based on multi-directional illumination image fusion technology is proposed. First, four grayscale images are taken with four bar-shape directional light sources from different directions. Next, based on the difference in surface grayscale characteristics for the different illumination directions of the surface’s stamped depression regions and flat regions, the image background is extracted and eliminated. Second, the images are fused using the difference processing on the images in the two groups of relative illuminant directions. Third, mean filter, binarization, and morphological closing operations are performed on the fused image to locate and segment the character string in the image, and the characters are normalized by correcting the skew of the segmented character string. Finally, histogram of oriented gradient features and a backpropagation neural network algorithm are employed to identify the normalized characters. Experimental results show that the algorithm can effectively eliminate the interference of factors such as oil stains, rust, oxide, shot-blasting pits, and different background colors and enhance the contrast between MSCs and background. The resulting character recognition rate can reach 99.6%.

FPGA architecture for text extraction from images

Image processing is a computationally intensive operation and it is typically done in software using CPU processing power that is readily available these days. The applications that make use of image processing require expensive and powerful CPUs to perform real-time operations. Hence a low cost FPGA based image processing solution becomes useful. This eliminates the need for powerful CPUs and at the same time real-time processing can be achieved relatively easier. This paper proposes an algorithm for extracting text information from images. With the help of FPGA, the pixels can be pipelined or processed in parallel in order to achieve increased processing speed. The gray scale conversion is done on the input image followed by image binarization. In order to extract text from input image, morphological close operation and connected component analysis are performed. The ultimate aim and motivation of this paper is detected and extract the words and letters from the input image with efficient hardware architecture. Simulation is done using VHDL coding and the analysis and synthesis results are carried out using the device XC7VX330T of Virtex7 family. FPGA provides higher flexibility since the architecture can be easily upgraded to meet the requirement. It is observed that area and power minimization is obtained by implementing an optimized design using this algorithm.

An efficient block-based algorithm for hair removal in dermoscopic images

Hair occlusion in dermoscopy images affects the diagnostic operation of the skin lesion. Segmentation and classification of skin lesions are two major steps of the diagnostic operation required by dermatologists. We propose a new algorithm for hair removal in dermoscopy images that includes two main stages: hair detection and inpainting. In hair detection, a morphological bottom-hat operation is implemented on Y-channel image of YIQ color space followed by a binarization operation. In inpainting, the repaired Y-channel is partitioned into 256 non-overlapped blocks and for each block, white pixels are replaced by locating the highest peak, using a histogram function and a morphological close operation. The proposed algorithm reports a true positive rate (sensitivity) of 97.36 %, a false positive rate (fall-out) of 4.25 %, and a true negative rate (specificity) of 95.75 %. The diagnostic accuracy achieved is recorded at a high level of 95.78 %.

Automatic Recognition of Sunspots in HSOS Full-Disk Solar Images

AbstractA procedure is introduced to recognise sunspots automatically in solar full-disk photosphere images obtained from Huairou Solar Observing Station, National Astronomical Observatories of China. The images are first pre-processed through Gaussian algorithm. Sunspots are then recognised by the morphological Bot-hat operation and Otsu threshold. Wrong selection of sunspots is eliminated by a criterion of sunspot properties. Besides, in order to calculate the sunspots areas and the solar centre, the solar limb is extracted by a procedure using morphological closing and erosion operations and setting an adaptive threshold. Results of sunspot recognition reveal that the number of the sunspots detected by our procedure has a quite good agreement with the manual method. The sunspot recognition rate is 95% and error rate is 1.2%. The sunspot areas calculated by our method have high correlation (95%) with the area data from the United States Air Force/National Oceanic and Atmospheric Administration (USAF/NOAA).

다중 이진화를 이용한 컨테이너 BIC 부호 영역 추출 및 인식 방법

The container BIC-code is a transport protocol for convenience in international shipping and combined transport environment. It is an identification code of a marine transport container which displays a wide variety of information including country's code. Recently, transportation through aircrafts and ships continues to rise. Thus fast and accurate processes are required in the ports to manage transportation. Accordingly, in this paper, we propose a BIC-code region extraction and recognition method using multiple thresholds. In the code recognition, applying a fixed threshold is not reasonable due to a variety of illumination conditions caused by change of weather, lightening, camera position, color of the container and so on. Thus, the proposed method selects the best recognition result at the final stage after applying multiple thresholds to recognition. For each threshold, we performs binarization, labeling, BIC-code pattern decision (horizontal or vertical pattern) by morphological close operation, and character separation from the BIC-code. Then, each characters is recognized by template matching. Finally we measure recognition confidence scores for all the thresholds and choose the best one. Experimental results show that the proposed method yields accurate recognition for the container BIC-code with robustness to illumination change. 키워드 : BIC-code, 이진화, 템플릿 매칭, 인식 Key word : BIC-code, threshold, template Matching, recognition

Automatic lung fields segmentation in CT scans using morphological operation and anatomical information

Segmenting lung fields from CT (Computed Tomography) scans is an important task for the analysis, diagnosis and treatment of pulmonary diseases. Although many segmentation methods have been presented, some new automatic segmentation methods for the lung fields are still proposed for the CT scans. This paper proposes a novel segmentation method for lung fields by using morphological closing operations and thresholding for normal lungs in CT scans. Additionally, under the guidance of anatomic information, the lung fields could be well segmented with lobar fissure, thin junction between the left/right lung fields, indentation of the blood vessels and bronchi-walls. This experiment is performed by employing the thoracic CT scans datasets, and it is proved to be an effective method.

Algorithms for morphological profile filters and their comparison

Morphological filters, regarded as the complement of mean-line based filters, are useful in the analysis of surface texture and the prediction of functional performance. The paper first recalls two existing algorithms, the naive algorithm and the motif combination algorithm, originally developed for the traditional envelope filter. With minor extension, they could be used to compute morphological filters. A recent novel approach based on the relationship between the alpha shape and morphological closing and opening operations is presented as well. Afterwards two novel algorithms are developed. By correlating the convex hull and morphological operations, the Graham scan algorithm, original developed for the convex hull is modified to compute the morphological envelopes. The alpha shape method depending on the Delaunay triangulation is costly and redundant for the computation for the alpha shape for a given radius. A recursive algorithm is proposed to solve this problem. A series of observations are presented for searching the contact points. Based on the proposed observations, the algorithm partitions the profile data into small segments and searches the contact points in a recursive manner. The paper proceeds to compare the five distinct algorithms in five aspects: algorithm verification, algorithm analysis, performance evaluation, end effects correction, and areal extension. By looking into these aspects, the merits and shortcomings of these algorithms are evaluated and compared.

Entropy-based feature extraction and decision tree induction for breast cancer diagnosis with standardized thermograph images

In this study, a computer-assisted entropy-based feature extraction and decision tree induction protocol for breast cancer diagnosis using thermograph images was proposed. First, Beier–Neely field morphing and linear affine transformation were applied in geometric standardization for whole body and partial region respectively. Gray levels of pixel population at the same anatomical position were statistically analyzed for abnormal region classification. Morphological closing and opening operations were used to identify unified abnormal regions. Three types of 25 feature parameters (i.e. 10 geometric, 7 topological and 8 thermal) were extracted for parametric factor analysis. Positive and negative abnormal regions were further reclassified by decision trees to induce the case-based diagnostic rules. Finally, anatomical organ matching was utilized to identify the corresponding organ with the positive abnormal regions. To verify the validity of the proposed cased-based diagnostic protocol, 71 and 131 female patients with and without breast cancer were analyzed. Experimental results indicated that 1750 abnormal regions (703 positive and 1047 negative) were detected and 822 branches were broken down into the decision space. Fourteen branches were found to have more than 4 positive abnormal regions. These critical diagnostic paths with less than 10% of positive abnormal regions (61/703 = 8.6%) can effectively classify more than half of the cancer patients (42/71 = 59.2%) in the abovementioned 14 branches.

Technique for Automated Recognition of Sunspots on Full-Disk Solar Images

A new robust technique is presented for automated identification of sunspots on full-disk white-light (WL) solar images obtained from SOHO/MDI instrument and Ca II K1 line images from the Meudon Observatory. Edge-detection methods are applied to find sunspot candidates followed by local thresholding using statistical properties of the region around sunspots. Possible initial oversegmentation of images is remedied with a median filter. The features are smoothed by using morphological closing operations and filled by applying watershed, followed by dilation operator to define regions of interest containing sunspots. A number of physical and geometrical parameters of detected sunspot features are extracted and stored in a relational database along with umbra-penumbra information in the form of pixel run-length data within a bounding rectangle. The detection results reveal very good agreement with the manual synoptic maps and a very high correlation with those produced manually by NOAA Observatory, USA.

Automated recognition of coronal mass ejections (CMEs) in near-real-time data

This paper presents a new method and first applications of software that we have developed to autonomously detect CMEs in image sequences from LASCO (Large Angle Spectrometric Coronagraph). The crux of the software is the detection of CMEs as bright ridges in (time, height) maps using the Hough transform. The next step employs clustering and morphological closing operations to mark out different CMEs. The output is a list of events, similar to the classic catalogs, with starting time, principle angle, angular width and velocity estimation for each CME. In addition we present a new type of CME overview map that clearly shows all detected CMEs in a (principal angle, time of occurrence) coordinate system. In contrast to catalogs assembled by human operators, these CME detections can be done without any human interference on real-time data 24 h per day (see http://sidc.oma.be/cactus for the real-time output with data covering the last 4 days). Therefore the detection is not only more immediate, but, more importantly, also more objective. In this paper we describe the software and validate its performance by comparing its output with the SOHO LASCO CME catalog. Experimental results on real-time data show that the developed technique can achieve excellent results in measuring starting time and principal angle and good results for the angular width and velocity measurement compared to the CMEs listed in the catalog. Its overall success rate is presently about 94%. The software also reveals CMEs or other features that have not been listed in the catalog. Such unreported cases might influence CME statistics and they demonstrate that also the present catalogs do not have a 100% success rate. This inevitably leads to a discussion on the definition of a CME. Prospects for improvement and exploitation are discussed.