2017_CAG_postprint.pdf - Google Drive
aCentre for Vision, Speech and Signal Processing (CVSSP), University of Surrey — Guildford, United Kingdom, GU2 7XH b Institute of Mathematical and Computer Sciences (ICMC), Universidade de S ̃ao Paulo — S ̃ao Carlos/SP, Brazil, 13566-590 A R T I C L E I N F O Article history: Received January 2, 2018 Keywords: Sketch based image retrieval (SBIR), Deep learning, Cross-domain modelling, Compact feature representa- tions, Multi-stage regression, Contrastive and triplet losses A B S T R A C T We propose and evaluate several deep network architectures for measuring the simi- larity between sketches and photographs, within the context of the sketch based image retrieval (SBIR) task. We study the ability of our networks to generalize across diverse object categories from limited training data, and explore in detail strategies for weight sharing, pre-processing, data augmentation and dimensionality reduction. In addition to a detailed comparative study of network configurations, we contribute
Explore this link on the map →