Context aware query image representation for particular object retrieval

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussavertaisarvioitu




The current models of image representation based on Convolutional Neural Networks (CNN) have shown tremendous performance in image retrieval. Such models are inspired by the information flow along the visual pathway in the human visual cortex. We propose that in the field of particular object retrieval, the process of extracting CNN representations from query images with a given region of interest (ROI) can also be modelled by taking inspiration from human vision. Particularly, we show that by making the CNN pay attention on the ROI while extracting query image representation leads to significant improvement over the baseline methods on challenging Oxford5k and Paris6k datasets. Furthermore, we propose an extension to a recently introduced encoding method for CNN representations, regional maximum activations of convolutions (R-MAC). The proposed extension weights the regional representations using a novel saliency measure prior to aggregation. This leads to further improvement in retrieval accuracy.


OtsikkoImage Analysis - 20th Scandinavian Conference, SCIA 2017, Proceedings
TilaJulkaistu - 2017
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisuussa
TapahtumaScandinavian Conference on Image Analysis - Tromso, Norja
Kesto: 12 kesäkuuta 201714 kesäkuuta 2017
Konferenssinumero: 20


NimiLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vuosikerta10270 LNCS
ISSN (painettu)03029743
ISSN (elektroninen)16113349


ConferenceScandinavian Conference on Image Analysis

ID: 14483862