Page 53 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

Conference Journal Journal Journal Conference Journal Conference Continued

[78] [79] [80] [50] [81] [82] [83]
Tumor a for as SWI
first children organized training to data set having optical 5378
the set second rolandic healthy Brain for method, images 320 Philips TBI images validation set 157 sets, taken
which data the a 17 proposed SWI referred included the MRIs, training of used ACHIKO-NC having data were

in and from including matched Multimodal Benchmark utilized of was 3.0T a using a into set was sets image SCES
used available BSIBSR 18 by were the built which SWI-CMB from done multichannel split and a process, data
were publicly the obtained study, and Segmentation 2013 testing of was set the System was images purpose, tomography used fundus and
sets a by is set (RE) epilepsy provided MICCAI as validation data detection, acquired evaluation 61 randomly 15 validation population-based were ORIGA
data is one provided data epilepsy with individuals sets Image well the large CMB SWI-CMB, images Medical with were having images evaluation coherence the images glaucoma namely,

Two Data by as For The set 46 For For The

coherence tomography imaging fundus imaging

MRI MRI MRI MRI Optical Slit-lamp Digital

CNN 3D CNN CNN

FCNN Multiscale CNN 3D Multiscale CNN Multiscale Convolutional- recursive neural network Deep

segmentation the of and detection of segmentation medical automatically the grading of diagnosis

the structures segmentation tumor automatic semantic from for cataracts the
for for brain for microbleeds for for prediction for features nuclear for
developed subcortical brain designed of designed cerebral developed lesion designed information developed the of developed

Model of human Model diagnosis Model of System brain Model image Model learning severity Model glaucoma

2017 2015

Eye