Systematic Evaluation of Convolution Neural Network Advances on the ImageNet

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F17%3A00311817" target="_blank" >RIV/68407700:21230/17:00311817 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.sciencedirect.com/science/article/pii/S1077314217300814" target="_blank" >http://www.sciencedirect.com/science/article/pii/S1077314217300814</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.cviu.2017.05.007" target="_blank" >10.1016/j.cviu.2017.05.007</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Systematic Evaluation of Convolution Neural Network Advances on the ImageNet

Popis výsledku v původním jazyce

The paper systematically studies the impact of a range of recent advances in convolution neural network (CNN) architectures and learning methods on the object categorization (ILSVRC) problem. The evaluation tests the influence of the following choices of the architecture: non-linearity (ReLU, ELU, maxout, compatability with batch normalization), pooling variants (stochastic, max, average, mixed), network width, classifier design (convolutional, fully-connected, SPP), image pre-processing, and of learning parameters: learning rate, batch size, cleanliness of the data, etc. The performance gains of the proposed modifications are first tested individually and then in combination. The sum of individual gains is greater than the observed improvement when all modifications are introduced, but the “deficit” is small suggesting independence of their benefits. We show that the use of 128 x 128 pixel images is sufficient to make qualitative conclusions about optimal network structure that hold for the full size Caffe and VGG nets. The results are obtained an order of magnitude faster than with the standard 224 pixel images.

Název v anglickém jazyce

Systematic Evaluation of Convolution Neural Network Advances on the ImageNet

Popis výsledku anglicky

The paper systematically studies the impact of a range of recent advances in convolution neural network (CNN) architectures and learning methods on the object categorization (ILSVRC) problem. The evaluation tests the influence of the following choices of the architecture: non-linearity (ReLU, ELU, maxout, compatability with batch normalization), pooling variants (stochastic, max, average, mixed), network width, classifier design (convolutional, fully-connected, SPP), image pre-processing, and of learning parameters: learning rate, batch size, cleanliness of the data, etc. The performance gains of the proposed modifications are first tested individually and then in combination. The sum of individual gains is greater than the observed improvement when all modifications are introduced, but the “deficit” is small suggesting independence of their benefits. We show that the use of 128 x 128 pixel images is sufficient to make qualitative conclusions about optimal network structure that hold for the full size Caffe and VGG nets. The results are obtained an order of magnitude faster than with the standard 224 pixel images.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

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

Projekt

<a href="/cs/project/GBP103%2F12%2FG084" target="_blank" >GBP103/12/G084: Centrum pro multi-modální interpretaci dat velkého rozsahu</a><br>

Návaznosti

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

Rok uplatnění

2017

Kód důvěrnosti údajů

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

Název periodika

Computer Vision and Image Understanding

ISSN

1077-3142

e-ISSN

1090-235X

Svazek periodika

161

Číslo periodika v rámci svazku

August

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

9

Strana od-do

11-19

Kód UT WoS článku

000410718600002

EID výsledku v databázi Scopus

2-s2.0-85019358051