16 Jul 2017
Paper accepted to ICCV 2017:
Look, Listen and Learn.
Camera-ready coming soon, in the mean time see the first version on arXiv.
11 Jul 2017
VGG Image Search Engine (VISE), mainly based on the codebase I developed during my PhD and PostDoc at VGG, has been open-sourced and released. It implements memory efficient large-scale retrieval using local descriptors, spatial reranking, Hamming embedding, query expansion, etc.
11 Jun 2017
Extended version of our CVPR 2016 NetVLAD paper accepted to TPAMI:
NetVLAD: CNN architecture for weakly supervised place recognition
23 May 2017
Paper published on arXiv:
Look, Listen and Learn
27 Feb 2017
Paper accepted to CVPR 2017 (spotlight):
Convolutional neural network architecture for geometric matching
11 Jan 2017
An extended version of our CVPR 2015 paper has been accepted to TPAMI:
24/7 place recognition by view synthesis
17 Sep 2016
If you missed our oral presentation of NetVLAD at CVPR 2016, now you can watch it on YouTube!
13 Sep 2016
Torsten Sattler has released code for his CVPR 2016 paper, which includes implementation of my ACCV 2014 work DisLocation: Scalable descriptor distinctiveness for location recognition, and more:
14 Jul 2016
Paper accepted to BMVC 2016:
Faces In Places: Compound query retrieval
04 Jul 2016
I have left the WILLOW group (INRIA / École Normale Superiéure) and am joining DeepMind as a Senior research scientist