ABSTRACT
Background: Smoothing in 99mTc myocardial perfusion single-photon emission computed tomography (SPECT) often increases myocardial artifacts due to subdiaphragmatic activity near the heart. To reduce these artifacts, we developed a new process flow, masking on unsmoothed images (MUS), that includes the extraction of the myocardium by masking before smoothing. Methods: This study evaluated the relationships between matrix sizes and distances to the subdiaphragmatic activity using the MUS method compared to conventional methods using a combination of image reconstruction methods (filtered back-projection [FBP] and ordered subset expectation maximization [OSEM]) with or without corrections (attenuation [AC], scatter [SC], and resolution recovery [RR]) using a myocardial phantom. The results were compared for two matrix sizes (pixel sizes) (128×128 [3.3 mm] and 64×64 [6.6 mm]); four subdiaphragmatic activity distances (5, 10, 15, and 20 mm); and three reconstruction methods (FBP without correction; OSEM with RR; and OSEM with AC, SC, and RR). Results: In the conventional method, increasing distance resulted in interference with myocardial perfusion SPECT evaluation however, the artifacts were less apparent when the MUS method was applied. The images converted to 64×64 did not show the same effect as the 128×128 images, even when RR was used. The MUS method was useful for acquisition at 128×128, along with the use of RR in the reconstruction process. Conclusion: MUS mitigated the effects of subdiaphragmatic activity on myocardial perfusion SPECT, particularly combined with 128×128 acquisitions and iterative reconstruction with RR.
ABSTRACT
This article proposes a methodology to forecast the movements of analysts' estimated net income and stock prices using analyst profiles. Our methodology is based on applying natural language processing and neural networks in the context of analyst reports. First, we apply the proposed method to extract opinion sentences from the analyst report while classifying the remaining parts as non-opinion sentences. Then, we employ the proposed method to forecast the movements of analysts' estimated net income and stock price by inputting the opinion and non-opinion sentences into separate neural networks. In addition to analyst reports, we input analyst profiles to the networks. As analyst profiles, we used the name of an analyst, the securities company to which the analyst belongs, the sector which the analyst covers, and the analyst ranking. Consequently, we obtain an indication that the analyst profile effectively improves the model forecasts. However, classifying analyst reports into opinion and non-opinion sentences is insignificant for the forecasts.
ABSTRACT
OBJECTIVE: We studied the behaviors of cost functions in the registration of thallium-201 (201Tl) brain tumor single-photon emission computed tomography (SPECT) and magnetic resonance (MR) images, as the similarity index of image positioning. METHODS: A marker for image registration [technetium-99m (99mTc) point source] was attached at three sites on the heads of 13 patients with brain tumor, from whom 42 sets of 99mTc-201Tl SPECT (the dual-isotope acquisition) and MR images were obtained. The 201Tl SPECT and MR images were manually registered according to the markers. From the positions where the two images were registered, the position of the 201Tl SPECT was moved to examine the behaviors of the three cost functions, i.e., ratio image uniformity (RIU), mutual information (MI), and normalized MI (NMI). RESULTS: The cost functions MI and NMI reached the maximum at positions adjacent to those where the SPECT and MR images were manually registered. As for the accuracy of image registration in terms of the cost functions MI and NMI, on average, the images were accurately registered within 3 degrees of rotation around the X-, Y-, and Z-axes, and within 1.5 mm (within 2 pixels), 3 mm (within 3 pixels), and 4 mm (within 1 slice) of translation to the X-, Y-, and Z-axes, respectively. In terms of rotation around the Z-axis, the cost function RIU reached the minimum at positions where the manual registration of the two images was substantially inadequate. CONCLUSIONS: The MI and NMI were suitable cost functions in the registration of 201Tl SPECT and MR images. The behavior of the RIU, in contrast, was unstable, being unsuitable as an index of image registration.
