Transcription of Abstract
{{id}} {{{paragraph}}}
Example: tourism industry
Abstract
deep Inside Convolutional Networks: VisualisingImage Classification Models and Saliency MapsKaren SimonyanAndrea VedaldiAndrew ZissermanVisual Geometry Group, University of paper addresses the visualisation of image classification models, learnt us-ing deep Convolutional Networks (ConvNets). We consider two visualisationtechniques, based on computing the gradient of the class score with respect tothe input image. The first one generates an image, which maximises the classscore [5], thus visualising the notion of the class, captured by a ConvNet. Thesecond technique computes a class saliency map, specific to a given image andclass. We show that such maps can be employed for weakly supervised objectsegmentation using classification ConvNets. Finally, we establish the connectionbetween the gradient-based ConvNet visualisation methods and deconvolutionalnetworks [13].
image space. The method was used to visualise the hidden feature layers of unsupervised deep ar-chitectures, such as the Deep Belief Network (DBN) [7], and it was later employed by Le et al.[9] to visualise the class models, captured by a deep unsupervised auto-encoder. Recently, the problem of ConvNet visualisation was addressed by Zeiler et ...
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: