Search

The manuscript conducts a comparative analysis to assess the impact of noise on medical images using a proposed threshold value estimation approach. It applies an innovative method for edge detection on images of varying complexity, considering different noise types and concentrations of noise. Five edges are evaluated on images with low, medium, and high detail levels. This study focuses on medical images from three distinct datasets: retinal images, brain tumor segmentation, and lung segmentation from CT scans. The importance of noise analysis is heightened in medical imaging, as noise can significantly obscure the critical features and potentially lead to misdiagnoses. Images are categorized based on the complexity, providing a multidimensional view of noise's effect on edge detection. The algorithm utilized the grid search (GS) method and random search with nine values (RS9). The results demonstrate the effectiveness of the proposed approach, especially when using the Canny operator, across diverse noise types and intensities. Laplace operators are most affected by noise, yet significant improvements are observed with the new approach, particularly when using the grid search method. The obtained results are compared with the most popular techniques for edge detection using deep learning like AlexNet, ResNet, VGGNet, MobileNetv2, and Inceptionv3. The paper presents the results via graphs and edge images, along with a detailed analysis of each operator's performance with noisy images using the proposed approach. [ABSTRACT FROM AUTHOR]

The influence of different wavelet transformations and decomposition on edge detection was examined, using convenient operators to images of various complexities. Berkeley Segmentation Database images with the corresponding ground truth were used. The categorization of those images was accomplished according to the degree of complexity in three groups (small, medium, and large number of details), by using discrete cosine transformation and discrete wavelet transformation. Three levels of decomposition for eight wavelet transformations and five operators for edge detection were applied on these images. As an objective measure of the quality for edge detection, the parameters "performance ratio" and "F-measure" were used. The obtained results showed that edge detection operators behaved differently in images with a different number of details. Decomposition significantly degrades the image, but useful information can be extracted at the third level of decomposition, because the image with a different number of details behaves differently at each level. For an image with a certain number of details, decomposition Level 3 in some cases gives better results than Level 2. The obtained results can be applied to image compression with different complexity. By selecting a certain combination of operators and decomposition levels, a higher compression ratio with preserving a larger amount of useful image information can be achieved. Depending on the image resolution whereby the number of details varies, an operator optimization can be performed according to the decomposition level in order to obtain the best possible edge detection. [ABSTRACT FROM AUTHOR]

This paper presents a new approach to edge detection based on the application of an adequate adaptive algorithm for edge detection depending on the image complexity criteria. The complexity of the image was determined based on spatial information. Also, a new approach of estimating threshold for edge detection has been developed, which is based on the principles of machine learning, by applying grid search and random search. For analysis and testing, images with real scenes were used. This dataset was used to improve threshold estimation using the grid and random search methods. Based on the obtained results, it can be seen that the complexity significantly affects the edge detection, so with a low and medium number of details, the Roberts operator achieved good results. For images with a high number of details, the Canny operator has proven to be the best solution. New approach for estimate the threshold based on the grid search improves edge detection, especially with the Canny operator. Grid search-based method gives best results but it takes a lot of computational time. If computational time and the improvements obtained are taken into account, more effective optimization of the parameters for estimating threshold values is random search method. [ABSTRACT FROM AUTHOR]