Explore chapters and articles related to this topic
A Novel Stacked Model Ensemble for Improved TB Detection in Chest Radiographs
Published in K.C. Santosh, Sameer Antani, D.S. Guru, Nilanjan Dey, Medical Imaging, 2019
Sivaramakrishnan Rajaraman, Sema Cemir, Zhiyun Xue, Philip Alderson, George Thoma, Sameer Antani
The datasets include PA CXRs that contain regions other than the lungs which are irrelevant for lung TB detection. To alleviate issues due to models learning features that are irrelevant to detecting lung TB and demonstrate sub-optimal performance, the lung region constituting the ROI is segmented by a method that uses anatomical atlases with non-rigid registration [24]. This segmentation method follows a content-based image retrieval approach to identify the training examples that bear resemblance to the patient CXR by using Bhattacharyya similarity measure and partial Radon transform. The patient-specific anatomical lung shape model is created using SIFT-flow [53] for registering the training masks for the patient CXRs. The refined lung boundaries are extracted using graph-cut optimization and customized energy function [54]. An instance of a CXR with the detected lung region and cropped lung area using the proposed method is shown in Figure 1.3.
Graph Cuts—Combinatorial Optimization in Vision
Published in Olivier Lézoray, Leo Grady, Image Processing and Analysis with Graphs, 2012
Move-making algorithms are iterative algorithms that repeatedly moves in the space of labelings so that the energy decreases. In general, the move can go to anywhere in the space, i.e., each site can change its label to any label; so, the best move would be to a global minimum. Since finding the global minimum in the general multi-label case is known to be NP-hard [106], we must restrict the range of possible moves so that finding the best move among them can be done in a polynomial time. The graph-cut optimization in each iteration of move-making algorithms can be thought of as prohibiting some of the labels at each site [107]. Depending on the restriction, this makes the optimization submodular or at least makes the graph much smaller than the full graph in the exact case.
Noise Removal Through the Exploration of Subjective and Apparent Denoised Patches Using Discrete Wavelet Transform
Published in IETE Journal of Research, 2021
K. Sakthidasan alias Sankaran, V. Nagarajan
The proposed image denoising technique utilizes the benefits of DWT and graph cut optimization. The block diagram of the proposed method is illustrated in Figure 1. The method is carried out in two stages namely, internal and external denoising. The two stages involved here includes four components such as registration, internal and external denoising and combining the two results. Image restoration is adopted to improve the correlation between the external and the noisy image. Graph cut based method is used in the first stage of external denoising and hard thresholding on transform coefficient of 3D cubes is used in the first stage of internal denoising. The Preliminary denoising result is produced at the first stage and its noise is assumed to be greatly decreased. This improves the denoising performance in the next stage.