Unsupervised online clustering and detection algorithms using crowdsourced data for malaria diagnosis. (February 2019)

• Record Type: Journal Article
• Authors: Pagès-Zamora, Alba; Cabrera-Bean, Margarita; Díaz-Vilor, Carles
• Is Part Of: Pattern recognition. Volume 86(2019:Feb.)
• Page Start: 209
• Page End: 223
• Publication Date: 2019-02
• Subjects: Crowdsourcing -- Unreliable annotators -- Unsupervised method -- Online EM algorithm -- MalariaSpot; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.09.001
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: An unsupervised method with crowdsourced data to detect forms in images is proposed. The procedure consists of a clustering and a detection stage based on the EM algorithm. The method accounts for outliers and is robust to unreliable annotators. An online implementation of the method suited for streaming data is presented. Experimental results with real data for Malaria diagnose support the approach. Abstract: Crowdsourced data in science might be severely error-prone due to the inexperience of annotators participating in the project. In this work, we present a procedure to detect specific structures in an image given tags provided by multiple annotators and collected through a crowdsourcing methodology. The procedure consists of two stages based on the Expectation–Maximization (EM) algorithm, one for clustering and the other one for detection, and it gracefully combines data coming from annotators with unknown reliability in an unsupervised manner. An online implementation of the approach is also presented that is well suited to crowdsourced streaming data. Comprehensive experimental results with real data from the MalariaSpot project are also included.