VGGNet
In plain wordsAn influential 2014 image-recognition network from Oxford, built by stacking many identical small pattern detectors, which showed that making networks deeper by itself improves accuracy.
A family of convolutional networksconvolutional neural network, n.: A kind of neural network built for images and other grid-like data. It slides small pattern detectors across the input, finding edges first, then shapes, then whole objects. with 11 to 19 weight layers, best known as VGG-16 and VGG-19, built from a strikingly uniform design: stacks of 3 × 3 convolutionsconvolution, n.: An operation that slides a small grid of weights across an image or signal, multiplying and adding at each spot, to measure how strongly a pattern appears at every location. with stridestride, n.: How far a scanning window jumps each time it moves across an image. Jumping two pixels instead of one halves the size of the result. 1 and padding 1, interleaved with 2 × 2 max-poolingpooling, n.: A step in an image network that shrinks its internal picture by summarizing each small patch with one number, usually the largest or the average, keeping what matters and discarding exact positions., with channels doubling from 64 to 512 as resolution halves, followed by three fully connected layersfully connected layer, n.: A layer in which every input is connected to every output by its own adjustable weight. It is the most general and most common kind of neural-network layer.. Two stacked 3 × 3 layers cover the same receptive fieldreceptive field, n.: The patch of the original input, such as a region of an image, that a particular neuron inside a network can see and respond to. as one 5 × 5 layer, and three cover 7 × 7, with fewer parameters and more nonlinearity, a principle later networks kept. VGG-16 has about 138 million parameters, most of them in its fully connected layers, so it is heavy to run. Its simplicity and the public release of trained weights made VGG features a standard choice for transfer learningtransfer learning, n.: Reusing what a model learned on one task to get a head start on another — like a pianist picking up the organ faster than a beginner. It lets models learn new jobs from little data., style transferstyle transfer, n.: Repainting a photo in the style of an artwork — your street as Van Gogh might have painted it — by keeping the photo's content and borrowing the painting's colors, strokes and textures. and perceptual losses long after more efficient designs overtook it.
The perceptual loss compares VGG-19 feature maps of the generated and target images.
Origin
Introduced by Karen Simonyan and Andrew Zisserman of the University of Oxford's Visual Geometry Group, which gave it its name, in 'Very Deep Convolutional Networks for Large-Scale Image Recognition' (arXiv, September 2014; ICLR 2015). Their entry placed first in localization and second in classification at the 2014 ImageNet challenge (ILSVRC).
Related
- convolutional neural networkconvolutional neural network, n.: A kind of neural network built for images and other grid-like data. It slides small pattern detectors across the input, finding edges first, then shapes, then whole objects.
- AlexNetAlexNet, prop. n.: The neural network that won a major image-recognition contest by a huge margin in 2012, convincing researchers and companies that deep learning trained on graphics chips was the way forward.
- InceptionInception, prop. n.: A family of Google image-recognition networks, the first called GoogLeNet, whose building blocks examine each patch of an image at several scales at once while keeping computation low.
- ResNetResNet, prop. n.: A family of image-recognition networks from 2015 whose shortcut connections allowed far deeper networks than before, over 100 layers, and became a standard building block across AI.
- receptive fieldreceptive field, n.: The patch of the original input, such as a region of an image, that a particular neuron inside a network can see and respond to.
- style transferstyle transfer, n.: Repainting a photo in the style of an artwork — your street as Van Gogh might have painted it — by keeping the photo's content and borrowing the painting's colors, strokes and textures.
Sources
- Simonyan, K. & Zisserman, A. (2014). Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv:1409.1556; ICLR 2015. arxiv.org/abs/1409.1556
Mentioned in
InceptionInception, prop. n.: A family of Google image-recognition networks, the first called GoogLeNet, whose building blocks examine each patch of an image at several scales at once while keeping computation low.