eprintid: 17287
rev_number: 14
eprint_status: archive
userid: 1336
dir: disk0/00/01/72/87
datestamp: 2014-09-11 07:35:08
lastmod: 2014-09-18 08:35:46
status_changed: 2014-09-11 07:35:08
type: doctoralThesis
metadata_visibility: show
creators_name: Wieler, Matthias
title: Multiple Instance Learning with Random Forests and Applications in Industrial Optical Inspection
subjects: 004
subjects: 310
subjects: 670
divisions: 130001
divisions: 708000
adv_faculty: af-13
abstract: Automatic defect detection in industrial optical inspection requires algorithms that can learn from data. A special challenge is data with incomplete labels. One of the methods that the field of machine learning has brought forth to deal with incomplete labels is multiple instance learning. One trait of this setting is that it groups datapoints (instances) into bags.

We propose a novel method to predict bag probabilities from given instance probabilities that has the advantage that its results do not depend on bag size. Also, we propose an extension of the multiple instance model that allows the user to steer the number of instances that are classified as positive.

We implement these methods with an algorithm based on the well-known random forest classifier. Results on a standard benchmark dataset show competitive performance. Furthermore, we apply this algorithm to image data that reflects the challenges of industrial optical inspection, and we show that in this setting it improves over the standard random forest.
date: 2014
id_scheme: DOI
id_number: 10.11588/heidok.00017287
ppn_swb: 1659197406
own_urn: urn:nbn:de:bsz:16-heidok-172875
date_accepted: 2014-07-23
advisor: HASH(0x564e1c2b4650)
language: eng
bibsort: WIELERMATTMULTIPLEIN2014
full_text_status: public
citation:   Wieler, Matthias  (2014) Multiple Instance Learning with Random Forests and Applications in Industrial Optical Inspection.  [Dissertation]     
document_url: https://archiv.ub.uni-heidelberg.de/volltextserver/17287/1/PhDThesis_Wieler.pdf