Unsupervised learning of spatial-temporal models of objects in a long-term autonomy scenario

Ambrus, RaresKTH,Datorseende och robotik, CVAP,Centrum för Autonoma System, CAS (author)

QC 20160411
We present a novel method for clustering segmented dynamic parts of indoor RGB-D scenes across repeated observations by performing an analysis of their spatial-temporal distributions. We segment areas of interest in the scene using scene differencing for change detection. We extend the Meta-Room method and evaluate the performance on a complex dataset acquired autonomously by a mobile robot over a period of 30 days. We use an initial clustering method to group the segmented parts based on appearance and shape, and we further combine the clusters we obtain by analyzing their spatial-temporal behaviors. We show that using the spatial-temporal information further increases the matching accuracy.

Ekekrantz, JohanKTH,Datorseende och robotik, CVAP,Centrum för Autonoma System, CAS(Swepub:kth)u1i6kh2t (author)
Folkesson, JohnKTH,Datorseende och robotik, CVAP,Centrum för Autonoma System, CAS(Swepub:kth)u1r1qd4z (author)
Jensfelt, PatricKTH,Datorseende och robotik, CVAP,Centrum för Autonoma System, CAS(Swepub:kth)u1u66xzd (author)
KTHDatorseende och robotik, CVAP (creator_code:org_t)

In:2015 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS): IEEE, s. 5678-56859781479999941

By the author/editor: Ambrus, Rares; Ekekrantz, Johan; Folkesson, John; Jensfelt, Patric

About the subject

NATURAL SCIENCES: NATURAL SCIENCES; and Computer and Inf ...; and Computer Vision ...

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.