In this paper, we develop a novel method for clustering multiple images while adjusting for the effect of available covariates on cluster membership. The key strategy is to represent each image as two-dimensional functional data and formulate a functional latent class mixed model, which fully leverages the structural information of images while effectively addressing their high dimensionality and accounting for covariate effects. We apply the proposed method to color intensity matrices extracted from patient-sourced smartphone fingernail photos to identify distinct subgroups, while adjusting for the effect of image metadata, which may act as an effect modifier on cluster membership. Information on these subgroups can assist public health officials in low-resource settings by enabling rapid and non-invasive identification of high-risk subpopulations for anemia, thereby facilitating the timely delivery of targeted interventions. The results suggest that the three clusters identified by the proposed method correspond to varying levels of anemia risk, with 0%, 79%, and 86% of subjects in each cluster classified as anemic. These findings demonstrate the utility of the proposed method and highlight the potential of a smartphone application leveraging fingernail images for non-invasive and cost-effective anemia screening.
WoldegebrielAGGebrehiwotGGDestaAA, et al.Identification of factors influencing anemia among children aged 6–59 months in Ethiopia using Ethiopia demographic and health survey 2016 data. Pediatric Health Med Ther2021; 12: 161–175.
2.
HailegebrealSNigatuAMMekonnenZA, et al.Spatio-temporal distribution and associated factors of anemia among children aged 6–59 months in ethiopia: A spatial and multilevel analysis based on the EDHS 2005–2016. BMJ Open2021; 11: e045544.
3.
ManninoRGMyersDRTyburskiEA, et al.Smartphone app for non-invasive detection of anemia using only patient-sourced photos. Nat Commun2018; 9: 1–10.
4.
TibshiraniR. Regression shrinkage and selection via the lasso. J R Stat Soc, B: Stat Methodol1996; 58: 267–288.
5.
KayanoMDozonoKKonishiS. Functional cluster analysis via orthonormalized Gaussian basis expansions and its application. J Classif2010; 27: 211–230.
6.
PengJMüllerHG, et al.Distance-based clustering of sparsely observed stochastic processes, with applications to online auctions. Ann Appl Stat2008; 2: 1056–1077.
7.
RossiFConan-GuezBEl GolliA. Clustering functional data with the SOM algorithm. In: Proceedings of the ESANN 2004, 2004, pp.305–312.
8.
AbrahamCCornillonPAMatzner-LøberE, et al.Unsupervised curve clustering using B-splines. Scand J Stat2003; 30: 581–595.
9.
IevaFPaganoniAMPigoliD, et al.Multivariate functional clustering for the morphological analysis of electrocardiograph curves. J R Stat Soc Ser, C: Appl Stat2013; 62: 401–418.
10.
SangalliLMSecchiPVantiniS, et al.K-mean alignment for curve clustering. Comput Stat Data Anal2010; 54: 1219–1233.
11.
TokushigeSYadohisaHInadaK. Crisp and fuzzy k-means clustering algorithms for multivariate functional data. Comput Stat2007; 22: 1–16.
12.
FerratyFVieuP. Nonparametric Functional Data Analysis: Theory and Practice. New York, NY: Springer, 2006, ISBN 978-0-387-36620-3.
YuPShiGWangC, et al.Distance-based clustering of functional data with derivative principal component analysis. J Comput Graph Stat2024; 34: 47–58.
15.
TangCShangHLYangY. Clustering and forecasting multiple functional time series. Ann Appl Stat2022; 16: 2523–2553.
16.
JacquesJPredaC. Functional data clustering: A survey. Adv Data Anal Classif2014; 8: 231–255.
17.
JacquesJPredaC. Funclust: A curves clustering method using functional random variables density approximation. Neurocomputing2013; 112: 164–171.
18.
BouveyronCJacquesJ. Model-based clustering of time series in group-specific functional subspaces. Adv Data Anal Classif2011; 5: 281–300.
19.
ChiouJMLiPL. Functional clustering and identifying substructures of longitudinal data. J R Stat Soc, B: Stat Methodol2007; 69: 679–699.
20.
CentofantiFLeporeAPalumboB. Sparse and smooth functional data clustering. Stat Pap2024; 65: 795–825.
21.
GuoWYouMYiJ, et al.Functional mixed effects clustering with application to longitudinal urologic chronic pelvic pain syndrome symptom data. J Am Stat Assoc2022; 117: 1631–1641.
22.
BouveyronCCômeEJacquesJ, et al.The discriminative functional mixture model for a comparative analysis of bike sharing systems. Ann Appl Stat2015; 9: 1726–1760.
23.
GiacofciMLambert-LacroixSMarotG, et al.Wavelet-based clustering for mixed-effects functional models in high dimension. Biometrics2013; 69: 31–40.
24.
SaméAChamroukhiFGovaertG, et al.Model-based clustering and segmentation of time series with changes in regime. Adv Data Anal Classif2011; 5: 301–321.
25.
HeardNAHolmesCCStephensDA. A quantitative study of gene regulation involved in the immune response of anopheline mosquitoes: An application of Bayesian hierarchical clustering of curves. J Am Stat Assoc2006; 101: 18–29.
26.
JamesGMSugarCA. Clustering for sparsely sampled functional data. J Am Stat Assoc2003; 98: 397–408.
27.
RigonT. An enriched mixture model for functional clustering. Appl Stoch Models Bus Ind2023; 39: 232–250.
28.
NguyenXGelfandAE. The dirichlet labeling process for clustering functional data. Stat Sin2011; 21: 1249–1289.
29.
PetroneSGuindaniMGelfandAE. Hybrid dirichlet mixture models for functional data. J R Stat Soc Ser, B: Stat Methodol2009; 71: 755–782.
30.
RaySMallickB. Functional clustering by Bayesian wavelet methods. J R Stat Soc, B: Stat Methodol2006; 68: 305–332.
31.
QuintanaFAMüllerPJaraA, et al.The dependent dirichlet process and related models. Stat Sci2022; 37: 24–41.
32.
DunsonDBPillaiNParkJH. Bayesian density regression. J R Stat Soc, B: Stat Methodol2007; 69: 163–183.
33.
ShiJQWangB. Curve prediction and clustering with mixtures of Gaussian process functional regression models. Stat Comput2008; 18: 267–283.
34.
WuHLiYF. Clustering spatially correlated functional data with multiple scalar covariates. IEEE T Neur Net Lear2022; 34: 7074–7088.
35.
SugarCAJamesGM. Finding the number of clusters in a dataset: An information-theoretic approach. J Am Stat Assoc2003; 98: 750–763.
36.
ReissTROgdenRT. Functional generalized linear modes with images as predictors. Biometrics2010; 66: 61–69.
37.
SaranliABaykalB. Complexity reduction in radial basis function (RBF) networks by using radial B-spline functions. Neurocomputing1998; 18: 183–194.
38.
HarrellFEHarrellF. Regression modeling strategies. Vol. 54. New York: Springer, 2001.
39.
MaiQZhangXPanY, et al.A doubly enhanced EM algorithm for model-based tensor clustering. J Am Stat Assoc2021; 117: 2120–2134.
40.
DwivediRHoNKhamaruK, et al.Singularity, misspecification and the convergence rate of EM. Ann Stat2020; 48: 3161–3182.
41.
HeinrichPKahnJ. Strong identifiability and optimal minimax rates for finite mixture estimation. Ann Stat2018; 46: 2844–2870.
42.
ChenJ. Optimal rate of convergence for finite mixture models. Ann Stat1995; 23: 221–233.
43.
FuWPerryPO. Estimating the number of clusters using cross-validation. J Comput Graph Stat2020; 29: 162–173.
44.
FujitaATakahashiDYPatriotaAG. A non-parametric method to estimate the number of clusters. Comput Stat Data Anal2014; 73: 27–39.
45.
FangYWangJ. Selection of the number of clusters via the bootstrap method. Comput Stat Data Anal2012; 56: 468–477.
46.
WangJ. Consistent selection of the number of clusters via crossvalidation. Biometrika2010; 97: 893–904.
47.
ChiangMMTMirkinB. Intelligent choice of the number of clusters in k-means clustering: An experimental study with different cluster spreads. J Classif2010; 27: 3–40.
48.
YanMYeK. Determining the number of clusters using the weighted gap statistic. Biometrics2007; 63: 1031–1037.
49.
FraleyCRafteryAE. Model-based clustering, discriminant analysis, and density estimation. J Am Stat Assoc2002; 97: 611–631.
50.
TibshiraniRWaltherGHastieT. Estimating the number of clusters in a data set via the gap statistic. J R Stat Soc, B: Stat Methodol2001; 63: 411–423.
51.
KrzanowskiWJLaiYT. A criterion for determining the number of groups in a data set using sum-of-squares clustering. Biometrics1988; 44: 23–34.
52.
MilliganGWCooperMC. An examination of procedures for determining the number of clusters in a data set. Psychometrika1985; 50: 159–179.
53.
HartiganJA. Clustering algorithms. New York, London, Sydney: John Wiley and Sons, 1975.
54.
CalińskiTHarabaszJ. A dendrite method for cluster analysis. Commun Stat - Theory Methods1974; 3: 1–27.
55.
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2023. https://www.R-project.org/.
KuhnHW. The Hungarian method for the assignment problem. Nav Res Logist1955; 2: 83–97.
59.
MunkresJ. Algorithms for the assignment and transportation problems. J Soc Ind Appl Math1957; 5: 32–38.
60.
GillespieARKahleABWalkerRE. Color enhancement of highly correlated images. I. Decorrelation and HSI contrast stretches. Remote Sens Environ1986; 20: 209–235.
