Research Article
Breast Microcalcification Diagnosis Using Deep Convolutional
Neural Network from Digital Mammograms
Hongmin Cai ,
1,2
Qinjian Huang,
1
Wentao Rong,
1
Yan Song,
1
Jiao Li,
3
Jinhua Wang,
4
Jiazhou Chen,
1
andLiLi
4
1
School of Computer Science and Engineering, South China University of Technology, Guangzhou 510000, China
2
Guangdong Provincial Key Lab of Computational Intelligence and Cyberspace Information,
South China University of Technology, Guangzhou, China
3
Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China,
Collaborative Innovation Center for Cancer Medicine, Guangzhou, Guangdong 510060, China
4
Medical Imaging Center, Shenzhen Hospital of Southern Medical University, Shenzhen, Guangdong 518101, China
Correspondence should be addressed to Hongmin Cai; hmcai@scut.edu.cn
Received 31 October 2018; Revised 22 December 2018; Accepted 4 February 2019; Published 3 March 2019
Academic Editor: Cristiana Corsi
Copyright©2019HongminCaietal.isisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Mammography is successfully used as an effective screening tool for cancer diagnosis. A calcification cluster on mam-
mography is a primary sign of cancer. Early researches have proved the diagnostic value of the calcification, yet their
performanceishighlydependentonhandcraftedimagedescriptors.Characterizingthecalcificationmammographyinan
automatic and robust way remains a challenge. In this paper, the calcification was characterized by descriptors obtained
fromdeeplearningandhandcrafteddescriptors.Wecomparedtheperformancesofdifferentimagefeaturesetsondigital
mammograms. e feature sets included the deep features alone, the handcrafted features, their combination, and
thefiltereddeepfeatures.Experimentalresultshavedemonstratedthatthedeepfeaturesoutperformhandcraftedfeatures,
but the handcrafted features can provide complementary information for deep features. We achieved a classification
precisionof89.32%andsensitivityof86.89%usingthefiltereddeepfeatures,whichisthebestperformanceamongallthe
feature sets.
1.Introduction
Breastcanceristhemostcommoncanceraffectingwomen’s
health. Early detection of breast cancer has been shown to
increase the survival rate, thereby significantly increasing
patients’ lifespan [1]. Mammography is a very popular
noninvasiveimagingtoolwithlowcostcomparedwithother
advanced equipment, such as computed tomography. It is
widelyusedtodiagnosebreastdiseaseatanearlystagedueto
its high sensitivity. erefore, it is frequently used as a tool
for early screening.
Duringmammographyscreening,thepresenceofbreast
microcalcifications (MCs) is a primary risk factor for breast
cancer. Breast calcifications in the early stages of breast
cancer appear like scattered spots in the mammographic
image that range from 0.1 to 1.0mm in size [2]. Previous
studies have found that MCs associated with malignant
lesionstendtobesmallerinsize,greaterinamount,andare
more densely distributed since they occur within the milk
ductsandotherassociatedstructuresinthebreastandfollow
theductalanatomy[3].Becauseahighcorrelationhasbeen
observed between the appearance of calcification clusters
and pathology results, the MCs provide a standard and
effective way for the automated detection of breast tumors.
Besides, large-scale genome-wide association studies
(GWAS) have proved to be a strong support for identifying
disease risk pathways [4]. Experimental results provide
clinicallyusefulcluesaboutthelinkbetweentheseriskgenes
and MS susceptibility in the Chinese population. e study
of [5] demonstrates convincingly that the genetic pre-
dispositionfordevelopmentofADisrootedintheimmune
system, rather than in neuronal cells in some degree.
Hindawi
Computational and Mathematical Methods in Medicine
Volume 2019, Article ID 2717454, 10 pages
https://doi.org/10.1155/2019/2717454