Purpose: To study, from a machine learning perspective, the performance of several machine learning classifiers that use texture analysis features extracted from soft-tissue tumors in nonenhanced T1-MRI images to discriminate between malignant and benign tumors. Results: The SVM classifier performs better than the neural network and the C4.5 decision tree based on the analysis of their receiver operating curves (ROC) and cost curves. The classification accuracy of the SVM, which was 93% (91% specificity; 94% sensitivity), was better than the radiologist classification accuracy of 90% (92% specificity; 81% sensitivity).
Conclusion:Machine learning classifiers trained with texture analysis features are potentially valuable for detecting malignant tumors in T1-MRI images. Analysis of the learning curves of the classifiers showed that a training data size smaller than 100 T1-MRI images is sufficient to train a machine learning classifier that performs as well as expert radiologists.
In this note, we address the estimation of the noise level in magnitude magnetic resonance (MR) images in the absence of background data. Most of the methods proposed earlier exploit the Rayleigh distributed background region in MR images to estimate the noise level. These methods, however, cannot be used for images where no background information is available. In this note, we propose two different approaches for noise level estimation in the absence of the image background. The first method is based on the local estimation of the noise variance using maximum likelihood estimation and the second method is based on the local estimation of the skewness of the magnitude data distribution. Experimental results on synthetic and real MR image datasets show that the proposed estimators accurately estimate the noise level in a magnitude MR image, even without background data.
Belgium. {jan. sijbers,dirk. vandyck}@ua. ac.be 3 UZ, Antwerpen, 2650 Edegem, Belgium, jan.gielen@uza.be S u m m a r y . Bias field signal is a low-frequency and very smooth signal that corrupts MRI images specially those produced by old MRI (Magnetic Resonance Imaging) machines. Image processing algorithms such as segmentation, texture analysis or classification that use the graylevel values of image pixels will not produce satisfactory results. A pre-processing step is needed to correct for the bias field signal before submitting corrupted MRI images to such algorithms or the algorithms should be modified. In this report we discuss two approaches to deal with bias field corruption. The first approach can be used as a preprocessing step where the corrupted MRI image is restored by dividing it by an estimated bias field signal using a surface fitting approach. The second approach shows how to modify the fuzzy c-means algorithm so that it can be used to segment an MRI image corrupted by a bias field signal.
Denoising of Magnetic Resonance images is important for proper visual analysis, accurate parameter estimation, and for further preprocessing of these images. Maximum Likelihood (ML) estimation methods were proved to be very effective in denoising Magnetic Resonance (MR) images. Among the ML based methods, the recently proposed Non Local Maximum Likelihood (NLML) approach gained much attention. In the NLML method, the samples for the ML estimation of the true underlying intensity are selected in a non local way based on the intensity similarity of the pixel neighborhoods. This similarity is generally measured using the Euclidean distance. A drawback of this approach is the usage of a fixed sample size for the ML estimation and, as a result, optimal results cannot be achieved because of over-or under-smoothing. In this work, we propose an NLML estimation method for denoising MR images in which the samples are selected in an adaptive way using the Kolmogorov-Smirnov (KS) similarity test. The method has been tested both on simulated and real data, showing its effectiveness.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.