Image Segmentation

Abstract

Image segmentation is one of the important and useful techniques in medical image processing. As the image segmentation technique results robust and high degree of accuracy, it is very much useful for the analysis of different image modalities, such as computerized tomography (CT) and magnetic resonance imaging (MRI) in the medical field. CT imaging gives more importance than MRI because of its wider availability, inexpensive and sensitiveness. In most cases, CT offers information needed to make decisions during urgent situations.

Development of magnetic image super-resolution model for nondestructive testing

“…We use two methods to evaluate the quality of images. Firstly, the structural similarity index measure (SSIM) [13] is used to measure the difference between structural information in the images rather than just evaluating the level of similarity in color. Therefore, SSIM gives results in the range from 0 to 1, with a value of 1 indicating perfect similarity between two images, while a negative value indicates that the two images are completely different.…”

“…In [11], a DL model with a Ushape structure was designed to nonlinearly map the input geometry of the MFLT system to the distribution of magnetic field around cracks. Single-image super-resolution, a fundamental challenge in computer vision, has received considerable attention, with researchers establishing end-to-end DNN structures to rebuild HRIs from LRIs [12][13][14][15]. Moreover, residual learning, a technique that preserves input data information in intermediate layers and forwards it to the output layer, has shown promise in addressing the issue of vanishing gradients during DNN training [16,17].…”

“…Notably, our proposed model effectively reduces noise in experimental data. To evaluate the performance of our approach, we utilized the structural similarity index measure (SSIM) [13] and root mean square error (RMSE) [14] metrics, comparing the real data (i.e. experimental data) with the data predicted by our DL model.…”

İnsan Embriyo Segmentasyonu için U-Net Tabanlı Modellerin Karşılaştırılması

“…Accuracy = 𝑇 𝑃 + 𝑇 𝑁 𝑇 𝑃 + 𝑇 𝑁 + 𝐹 𝑃 + 𝐹 𝑁 (1) The Jaccard index, also called intersection over union (IOU), is a similarity measure that evaluates the number of pixels shared between the ground truth and prediction masks divided by their union.…”

“…Image segmentation separates images into coherent regions based on deterministic features such as color, intensity value, and texture of pixels in the image and is usually the first step of image analysis. The main purpose of image segmentation in medical image processing applications is to optimize disease diagnosis by detecting the required region of interest (ROI) using an automated tool or algorithm [1]. Current studies show that although image segmentation is not a simple process, it is an essential step for diagnosing disease and isolating the ROI in different medical imaging modalities [2].…”