China
Africa
Explore chapters and articles related to this topic
Swarm Intelligence-based Framework for Image Segmentation of Knee MRI Images for Detection of Bone Cancer
Published in Shikha Agrawal, Manish Gupta, Jitendra Agrawal, Dac-Nhuong Le, Kamlesh Kumar Gupta, Swarm Intelligence and Machine Learning, 2022
Sujatha Jamuna Anand, C Tamilselvi, Dahlia Sam, C Kamatchi, Nibedita Dey, K. Sujatha
Image processing has been used in many scientific fields, such as in medicine or biology, however the researchers represent different types of cells by their texture properties [6], or distinguish between the alive or dead cells by analysing their images [7]. Edge detection is an important operation in image processing, which reduces the number of pixels and saves the image structure by determining the boundaries of objects in the it. The two general approaches to edge detection that are commonly used are gradient and Laplacian. The gradient method uses the first derivative of the image, and the Laplacian method uses the second derivative of the image to find edges. Pre-processing of an image is performed for the improvement of the image data and also for identifying image features which are important for further processing.
Convolution and Image Filtering
Published in Yevgeniy V. Galperin, An Image Processing Tour of College Mathematics, 2021
Sometimes, however, the exact opposite needs to be done - something tantamount to turning a painting into a drawing. Edge detection is a procedure for finding the boundaries of objects within images. It is used in image processing for the purposes of image segmentation, feature detection, and feature extraction. Edge detection has numerous applications in various areas such as computer and machine vision,video surveillance,satellite imaging,medicine,geology,glaciology,
Segmentation Techniques
Published in Jyotismita Chaki, Nilanjan Dey, A Beginner's Guide to Image Preprocessing Techniques, 2018
Jyotismita Chaki, Nilanjan Dey
Edge segmentation is a vital area of research, as it helps higher-level image exploration [15]. Detection of edges is an important tool for image segmentation. The representation of the edge of an image meaningfully decreases the amount of data to be processed, however, it holds vital information about the shapes of objects in the scene [16]. Edges are basically local variations in image intensity. Edge detection approaches convert original images into edge images depending on the variations of gray tones in the image. Image edge detection is used in many applications like object shape identification, medical image processing, biometrics, and so on [17,18]. There are three different types of discontinuities in the gray level such as points, lines, and edges. Spatial masks can be used to identify these three types of image discontinuities.
Detection and classification of pavement damages using wavelet scattering transform, fractal dimension by box-counting method and machine learning algorithms
Published in Road Materials and Pavement Design, 2023
Lizette Tello-Cifuentes, Johannio Marulanda, Peter Thomson
Edge detection is one of the most important basic features in image processing. Edges are made up of a set of pixels that define an abrupt variation and reflect discontinuity. The Prewitt edge detector and the Canny edge detector, two methods that have been successfully applied for edge detection in previous work by other researchers, were used in the methodology. The Prewitt edge detector obtains the edges of the image by convolution; and operates utilising directional filters, which are used to calculate horizontal and vertical edges (Li et al., 2015). Canny's method applies a Gaussian convolution to the image to smooth it, then calculates the intensity discontinuities in the image using the first derivative; producing an intensity of edge and direction in each pixel of the smoothed image (Hoang & Nguyen, 2018a Muduli & Pati, 2013;). All images were preprocessed using both edge detection methods. Results show that the Canny method is more effective in identifying longitudinal cracks because it generates better edge continuity (Figure 4), while the Prewitt method provides better results for pothole detection and alligator cracks.
Infrared Image Edge Recognition Algorithm Based on Partial Differential Equation for Industry 4.0
Published in IETE Journal of Research, 2022
IR image edge detection is one of the most powerful tools for improving and detecting edges in images. Identifying and locating edges is indeed a low-level task in various applications, such as 3-D reconstruction, shape recognition, image compression, enhancement, and restoration. But this is a crucial task. A discontinuity in brightness is detected. In image processing, computer vision, and machine vision, edge detection is used for image segmentation and data extraction. While infrared imaging radar range image feature recognition is being performed, it will be affected by the outside noise, which will decrease the accuracy of classification. To process a range image with rich contrast, the threshold of two images is obtained, and then the image is multiplied. Finally, the modulation is de-noised.
Retinal blood vessels segmentation using Wald PDF and MSMO operator
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
Sushil Kumar Saroj, Rakesh Kumar, Nagendra Pratap Singh
Sobel, Canny, Gradient, and Prewitt operators are the classical methods for edge detection. They work well when edges are distinct and sharp. Machine learning methods such as supervised algorithms need a basic understanding of labelling for the identification of vessel as well as non-vessel pixels. Unsupervised algorithms look at inherent features of vessels for the detection of a vessel as well as non-vessel pixels. Deep learning methods require bulk datasets to get output better than other methods. These methods are very costly for training because of their sophisticated data-models. Vessel tracking, morphology, multiscale, model based and kernel-based methods are rule-based methods. Kernel-based method is also called matched filter method. Matched filter method is assumed best to extract vessels as it enhances the contrast of vessels. Matched filter methods enhance the features of vessels by rotating an image with designed two-dimensional templates (kernels). Templates are formed to simulate the assigned vessel feature patterns and rotation results refer to the presence of the pattern. In matched filter methods, Gaussian probability distribution function (PDF) based matched filter is the first matched filter for vessel extraction. This method often detects non-vessels as vessels. Like Gaussian, Cauchy, and Gamma PDF (Kumar et al. 2021) based matched filters also do not match well with the intensity profile of a retinal vessel. Hence, the segmentation accuracy of these methods is low. Weibull, Gumbel, and Fréchet PDFs are skewed PDFs. Matched filter kernels based on these PDFs also do not match well with the intensity profile of a vessel.