Local Features and Visual Words Emerge in Activations

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F19%3A00337163" target="_blank" >RIV/68407700:21230/19:00337163 - isvavai.cz</a>

Result on the web

<a href="https://ieeexplore.ieee.org/document/8954470/keywords#keywords" target="_blank" >https://ieeexplore.ieee.org/document/8954470/keywords#keywords</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/CVPR.2019.01192" target="_blank" >10.1109/CVPR.2019.01192</a>

Result language

angličtina

Original language name

Local Features and Visual Words Emerge in Activations

Original language description

We propose a novel method of deep spatial matching (DSM) for image retrieval. Initial ranking is based on image descriptors extracted from convolutional neural network activations by global pooling, as in recent state-of-the-art work. However, the same sparse 3D activation tensor is also approximated by a collection of local features. These local features are then robustly matched to approximate the optimal alignment of the tensors. This happens without any network modification, additional layers or training. No local feature detection happens on the original image. No local feature descriptors and no visual vocabulary are needed throughout the whole process. We experimentally show that the proposed method achieves the state-of-the-art performance on standard benchmarks across different network architectures and different global pooling methods. The highest gain in performance is achieved when diffusion on the nearest-neighbor graph of global descriptors is initiated from spatially verified images.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

Result was created during the realization of more than one project. More information in the Projects tab.

Continuities

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Publication year

2019

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

CVPR 2019: Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition

ISBN

978-1-7281-3293-8

ISSN

1063-6919

e-ISSN

2575-7075

Number of pages

10

Pages from-to

11643-11652

Publisher name

IEEE

Place of publication

—

Event location

Long Beach

Event date

Jun 15, 2019

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—