Gp-Unet: Lesion detection from weak labels with a 3D regression network

Florian Dubost, Gerda Bortsova, Hieab H. Adams, M. Arfan Ikram, Wiro J. Niessen, Meike W. Vernooij, Marleen de Bruijne

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

20 Citations (Scopus)


We propose a novel convolutional neural network for lesion detection from weak labels. Only a single, global label per image - the lesion count - is needed for training. We train a regression network with a fully convolutional architecture combined with a global pooling layer to aggregate the 3D output into a scalar indicating the lesion count. When testing on unseen images, we first run the network to estimate the number of lesions. Then we remove the global pooling layer to compute localization maps of the size of the input image. We evaluate the proposed network on the detection of enlarged perivascular spaces in the basal ganglia in MRI. Our method achieves a sensitivity of 62% with on average 1.5 false positives per image. Compared with four other approaches based on intensity thresholding, saliency and class maps, our method has a 20% higher sensitivity.

Original languageEnglish
Title of host publicationMedical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
EditorsM. Descoteaux
Number of pages8
Volume10435 LNCS
ISBN (Print)9783319661780
Publication statusPublished - 2017
EventMedical Image Computing and Computer-Assisted Intervention, MICCAI 2017: 20th International Conference - Quebec City, Canada
Duration: 11 Sep 201713 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10435 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349


ConferenceMedical Image Computing and Computer-Assisted Intervention, MICCAI 2017
CityQuebec City

Fingerprint Dive into the research topics of 'Gp-Unet: Lesion detection from weak labels with a 3D regression network'. Together they form a unique fingerprint.

Cite this