China
Africa
Explore chapters and articles related to this topic
Algorithms for Drawing Graphics Primitives on a Honeycomb Model-Inspired Grid
Published in D. P. Acharjya, V. Santhi, Bio-Inspired Computing for Image and Video Processing, 2018
Since a hexagonal grid offers greater angular resolution, the nearest neighboring pixels of the same type are separated by 60∘ $ 60^{\circ } $ instead of 90∘ $ 90^{\circ } $ . In a square grid, pixels often disturb the perception of continuity of a circle. But due to better pixel connectivity in a hexagonal grid, circles are approximated by tiny polylines. The brightness of the circle is also found to be constant along its circumference (the is not so in a square grid) because of equidistant pixel neighbors. These result in circles of better quality than that in a square grid. In particular, the vertical and horizontal regions of the circles drawn by a square grid algorithm appear as a thick straight line. The overall shape of the circles, as the radii are reduced, worsens in the square lattice compared to the hexagonal lattice. This is because of the improved angular resolution afforded by having six equidistant neighboring pixels compared to only four in a square lattice. This is especially noticeable when the radius of the circle is 1. The oblique effect [45,21] also plays a part in this perception.
Biologically motivated approach to multistage image processing
Published in Waldemar Wójcik, Andrzej Smolarz, Information Technology in Medical Diagnostics, 2017
L.I. Timchenko, N.I. Kokryatskaya, I.D. Ivasyuk, T. Małecka-Massalska, Z. Omiotek, B. Akhmetov
When using the multistage approach for the representation of structural information within the image, there is no need to reconstruct the initial input from the obtained principal components. The main goal here is to discover some spatial regularities in order to interpret the image structure. However, reconstruction of the raw image is necessary in some applications, such as image coding and filtration. To apply the multistage approach to these tasks, information about the spatial transformations must be stored at each processing stage. Several transformations can be applied in the multistage network. Some possible examples include principal component analysis of local image patches (Rao & Ballard 1995), partial clusterisation of the Gabor coefficients (Daugman 1985), and the cosine transform. In this paper, analysis of the facial structure is performed using partial clusterisation of pixel connectivity. A three-level representation (White & Rohrer 1983, Kutaev 1989) of grey-scale facial image is used to specify a three-level pixel connectivity. The three connectivity levels are then processed using three multistage networks, each producing a pattern vector (Fig. 9.2).
Swarm Optimization and Machine Learning to Improve the Detection of Brain Tumor
Published in Shikha Agrawal, Manish Gupta, Jitendra Agrawal, Dac-Nhuong Le, Kamlesh Kumar Gupta, Swarm Intelligence and Machine Learning, 2022
A segmentation method based on ABC has been discussed in [28] for extracting tumors from the MRI brain images. The whole process goes through three phases: preprocessing, processing and post processing. In the preprocessing stage, noise is removed by using filters like average, median and Guassian Adaptive filter. In the processing stage the ABC algorithm is used for segmentation of the images. The image is converted into binary images by using gray level thresholding. The tumor region is then extracted from the segmented image by using connected component labeling on the basis of pixel connectivity. The results obtained by this algorithm are better as compared to other algorithms based on K-means and FCM as they fail to separate tumor regions from other parts.
A two-step sequential automated crack detection and severity classification process for asphalt pavements
Published in International Journal of Pavement Engineering, 2022
Thai Son Tran, Van Phuc Tran, Hyun Jong Lee, Julius Marvin Flores, Van Phuc Le
Image processing is a common technique used in analysing, evaluating, and extracting useful information from images. To identify objects in an image using image processing, pixels of the same colour that are connected must be determined. This is done by determining its pixel connectivity which is the relationship between pixels and its neighbours in 2-dimension as shown in Figure 4. The image is first transformed to a binary image in which the background is given a pixel value of 0 while the object is 255. For an N×N sized binary image, the pixel at the coordinate (x, y), where 0 ≤ x ≤ N − 1 and 0 ≤ y ≤ N − 1, is denoted as p (x, y). The definition of pixel connectivity component is that the pixel p(x, y) should be connected into any of its adjacent pixels (He et al. 2017b).