Journal: Neuron

Article Title: Direction selectivity in retinal bipolar cell axon terminals

doi: 10.1016/j.neuron.2021.07.008

Figure Lengend Snippet:

Article Snippet: The stack of adjusted images was filtered with a Gaussian filter (3 × 3 pixels), and then each image was downsampled to 0.8 of the original using a MATLAB imresize function.

Techniques: Virus, Plasmid Preparation, Recombinant, Software, Imaging

Journal: Journal of Applied Clinical Medical Physics

Article Title: Super‐resolution of brain tumor MRI images based on deep learning

doi: 10.1002/acm2.13758

Figure Lengend Snippet: Comparison of bicubic, overcomplete dictionaries, MRBT‐SR‐without perceptual loss, MRBT‐SR‐with perceptual loss on benchmark data

Article Snippet: Each input axial slice of an MRI T2 FLAIR image was normalized through the following steps: (1) the mean intensity value and the standard deviation of the foreground pixels were calculated, (2) the intensity value was subtracted by mean intensity value, and then divided by the standard deviation value for each pixel (including the background pixels), and (3) the high‐resolution normalized images were downsampled by a scaling factor of four using the MATLAB bicubic kernel function.

Techniques: Comparison

Journal: Journal of Applied Clinical Medical Physics

Article Title: Super‐resolution of brain tumor MRI images based on deep learning

doi: 10.1002/acm2.13758

Figure Lengend Snippet: Results of super‐resolution methods: (a) 4× downsampling of the original MRI image, (b) bicubic upsampling, (c) overcomplete dictionaries, (d) enhanced super‐resolution generative adversarial networks, (e) MRI‐based brain tumor super‐resolution (MRBT‐SR) with visual geometry group perceptual loss, (f) MRBT‐SR without perceptual loss, (g) MRBT‐SR with perceptual loss (Stage 1), (h) MRBT‐SR with perceptual loss (Stage 2), (i) MRBT‐SR with perceptual loss (Stage 3), (j) MRBT‐SR with perceptual loss (Stage 4), (k) the original high‐resolution image

Article Snippet: Each input axial slice of an MRI T2 FLAIR image was normalized through the following steps: (1) the mean intensity value and the standard deviation of the foreground pixels were calculated, (2) the intensity value was subtracted by mean intensity value, and then divided by the standard deviation value for each pixel (including the background pixels), and (3) the high‐resolution normalized images were downsampled by a scaling factor of four using the MATLAB bicubic kernel function.

Techniques:

Journal: Journal of Applied Clinical Medical Physics

Article Title: Super‐resolution of brain tumor MRI images based on deep learning

doi: 10.1002/acm2.13758

Figure Lengend Snippet: Comparison of improved performance contributed to brain tumor segmentation using different super‐resolution methods

Article Snippet: Each input axial slice of an MRI T2 FLAIR image was normalized through the following steps: (1) the mean intensity value and the standard deviation of the foreground pixels were calculated, (2) the intensity value was subtracted by mean intensity value, and then divided by the standard deviation value for each pixel (including the background pixels), and (3) the high‐resolution normalized images were downsampled by a scaling factor of four using the MATLAB bicubic kernel function.

Techniques: Comparison