Very Deep Convolutional Networks for Large-Scale Image Recognition

E366102
computer vision paper research paper scientific publication

"Very Deep Convolutional Networks for Large-Scale Image Recognition" is the influential 2014 research paper that introduced the VGG family of deep convolutional neural network architectures, demonstrating that significantly increasing network depth with small convolutional filters leads to substantial improvements in image classification performance.

Try in SPARQL Jump to: Surface forms Statements Referenced by

All labels observed (3)

Label Occurrences
Very Deep Convolutional Networks for Large-Scale Image Recognition canonical 4
VGG paper 1
VGG-16 and VGG-19 achieve state-of-the-art performance on ImageNet at the time of publication 1

Statements (46)

Predicate Object
instanceOf computer vision paper ⓘ
research paper ⓘ
scientific publication ⓘ
application image feature extraction ⓘ
object recognition ⓘ
transfer learning ⓘ
architectureStyle sequential convolutional layers ⓘ
very deep convolutional neural network ⓘ
benchmark ImageNet ⓘ
surface form: ILSVRC-2014

ImageNet ⓘ
category convolutional neural networks ⓘ
datasetSize over one million training images ⓘ
demonstrates benefits of using small 3x3 filters instead of larger filters ⓘ
effectiveness of depth over width in CNN design ⓘ
designPrinciple homogeneous architecture with similar layer configurations ⓘ
use of small convolutional filters stacked to increase receptive field ⓘ
domain artificial intelligence ⓘ
evaluationMetric top-1 error ⓘ
top-5 error ⓘ
field computer vision ⓘ
deep learning ⓘ
machine learning ⓘ
impact highly influential in the development of deep CNN architectures ⓘ
widely adopted as a baseline model in computer vision research ⓘ
influenced later CNN architectures such as ResNet ⓘ
use of VGG-style backbones in many vision models ⓘ
inspired pretrained CNN feature extraction in many applications ⓘ
use of deep feature representations for downstream tasks ⓘ
language English ⓘ
lossFunction softmax cross-entropy ⓘ
mainContribution demonstration that increasing network depth with small convolution filters improves image classification performance ⓘ
introduction of the VGG family of deep convolutional neural networks ⓘ
optimizationMethod stochastic gradient descent ⓘ
proposes VGG ⓘ
surface form: VGG-11 architecture

VGG ⓘ
surface form: VGG-13 architecture

VGG ⓘ
surface form: VGG-16 architecture

VGG ⓘ
surface form: VGG-19 architecture
shortTitle Very Deep Convolutional Networks for Large-Scale Image Recognition self-linksurface differs ⓘ
surface form: VGG paper
shows Very Deep Convolutional Networks for Large-Scale Image Recognition self-linksurface differs ⓘ
surface form: VGG-16 and VGG-19 achieve state-of-the-art performance on ImageNet at the time of publication

deeper convolutional networks can achieve better accuracy than shallower ones ⓘ
task large-scale image classification ⓘ
title Very Deep Convolutional Networks for Large-Scale Image Recognition self-link ⓘ
uses 1x1 convolutional filters ⓘ
3x3 convolutional filters ⓘ
max pooling layers ⓘ
year 2014 ⓘ

Referenced by (6)

Full triples — surface form annotated when it differs from this entity's canonical label.

VGG → describedInPaper → Very Deep Convolutional Networks for Large-Scale Image Recognition ⓘ
Karen Simonyan → notableWork → Very Deep Convolutional Networks for Large-Scale Image Recognition ⓘ
Visual Geometry Group → notableWork → Very Deep Convolutional Networks for Large-Scale Image Recognition ⓘ
Very Deep Convolutional Networks for Large-Scale Image Recognition → title → Very Deep Convolutional Networks for Large-Scale Image Recognition self-link ⓘ
Very Deep Convolutional Networks for Large-Scale Image Recognition → shortTitle → Very Deep Convolutional Networks for Large-Scale Image Recognition self-linksurface differs ⓘ
this entity surface form: VGG paper
Very Deep Convolutional Networks for Large-Scale Image Recognition → shows → Very Deep Convolutional Networks for Large-Scale Image Recognition self-linksurface differs ⓘ
this entity surface form: VGG-16 and VGG-19 achieve state-of-the-art performance on ImageNet at the time of publication