Page 57 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

Journal Journal Journal Journal Conference Journal Continued

[103] [104] [105] [106] [107] [108]

at Francisco July sets of total research X-ray for and (LIDC- used, Dataset (FMD), KTH-TIPS-
obtained San and data Institute 10,848 of set data the adult 35,038 radiographs chest Diagnostic Medical used was Consortium Initiative study the were Textures Database Group's

patients California, 2013 CXR Korean total of consisting Shenzhen images by approved of chest frontal 637 the Sheba of Israel, Resource for used sets data of library Textures (KTB), Research

909 January deidentified namely, consisting set 662 consisting of from Hashomer, Database was Describable Material Database (UIUC)
from of between used data MC and total of was consisting collected Department Image Database set benchmark Amsterdam Flickr Texture Ponce the database
images University (UCSF) were different used, Tuberculosis images, images, consisting used set board posterioreanterior set images, Imaging Tel Center, study Lung data texture namely, the (ALOT), the Kylberg and Texture

1885 the 2015 Three were 138 Data ethics Data the The Image IDRI) Six (DTD), 2b,

radiography X-rays radiography X-rays

Chest Chest (CXR) Chest Chest CT CT

CNN CNN CNN CNNdmultiple instance learning
Deep CNN Deep Decaf CNN

high X- for diagnosis
of with of of feasibility chest a that features malignancy
classification images a construct segmentation semantic computer-aided detection chest frontal the test in formulate framework deep cancer early the diseases

for relevant to for supervised for and on to pathology to transferable lung for lung
designed designed framework developed classification abnormalities radiographs developed detecting designed domain-adaptation patient-level prediction developed interstitial

Model clinically fidelity Model weakly Model System of rays System learns Model of