|
Title:
|
Discovering routine in egocentric images
|
|
Author:
|
Wuerich, Carolin
|
|
Other authors:
|
Universitat Politècnica de Catalunya; Radeva, Petia |
|
Abstract:
|
We propose a new unsupervised approach for the discovery of routine from egocentric photostreams. The method integrates Convolutional Neural Networks for image feature extraction with Latent Dirichlet Allocation to find topics during the days. Finally, the days are clustered based on their distances. |
|
Subject(s):
|
-Àrees temàtiques de la UPC::Informàtica
-Neural networks (Computer science)
-Imaging systems
-egocentric photostreams
-wearable camera
-visual life-logging
-Spectral Clustering
-routine discovery
-convolutional neural networks
-latent dirichlet allocation
-dynamic time warping
-spectral clustering
-Xarxes neuronals (Informàtica)
-Imatgeria (Tècnica) |
|
Rights:
|
|
|
Document type:
|
Research/Master Thesis |
|
Published by:
|
Universitat Politècnica de Catalunya
|
|
Share:
|
|
