GPI - Grup de Processament d'Imatge i Vídeo
El GPI fa recerca en Processament d'Imatge i Vídeo per representació, codificació, indexació i anàlisi del contingut visual. L'expertesa del grup en Morfologia i segmentació ha estat la base de contribucions als estàndards ISO MPEG-4 i MPEG-7. La recerca en anàlisi d’imatge l'ha permès participar en projectes europeus des de 1992 als programes RACE (Morpheco, coord.), ACTS (MAVT, MoMuSys, Vidas), IST (Diceman, Hypermedia, INTERFACE, ADViSOR, MASCOT, FAETHON), xarxes d'excel•lència (SCHEMA, SIMILAR, MUSCLE), i projectes integrats FP6 (CHIL) i FP7 (FASCINATE). El grup ha construit dues smart rooms al Campus Nord de la UPC, i ha fet contribucions en anàlisi visual per interacció, així com en aplicacions d’imatge biomèdica i teledetecció. Ha signat convenis de recerca amb empreses com Philips (París), France Telecom (Rennes), NXP (Holanda), Thomson (Princeton, USA), Alterface (Bèlgica) i nacionals com Telefònica, CCRTV, MediaPro, Fundació CELLEX, Hospital Clínic, AD Telecom o Abertis.
Collections
-
Articles de revista [75]
-
Capítols de llibre [12]
-
Llibres [2]
-
Reports de recerca [15]
Recent Submissions
-
Skip RNN: learning to skip state updates in recurrent neural networks
(2018)
Conference lecture
Open AccessRecurrent Neural Networks (RNNs) continue to show outstanding performance in sequence modeling tasks. However, training RNNs on long sequences often face challenges like slow inference, vanishing gradients and difficulty ... -
Foreground objects segmentation for moving camera scenarios based on SCGMM
(2011)
Conference report
Restricted access - publisher's policyIn this paper we present a new system for segmenting non-rigid objects in moving camera sequences for indoor and outdoor scenarios that achieves a correct object segmentation via global MAP-MRF framework formulation for ... -
Motion analysis of image sequences using connected operators
(International Society for Photo-Optical Instrumentation Engineers (SPIE), 1997)
Conference report
Restricted access - publisher's policyThis paper deals with a class of morphological operators called connected operators. These operators interact with the signal by merging flat zones. As a results, they do not create any new contours and are very attractive ... -
Active mesh coding and rate-distortion theory
(Institute of Electrical and Electronics Engineers (IEEE), 1996)
Conference lecture
Open AccessThis paper presents a video coding scheme for very low bit rate applications. The coding approach relies on active meshes and can be viewed as a particular case of region-based coding. The active mesh is used to efficiently ... -
Report on interim demonstration. FascinatE deliverable D6.2.1
(2012-07-25)
External research report
Open Access -
SaltiNet: scan-path prediction on 360 degree images using saliency volumes
(IEEE Press, 2018)
Conference lecture
Open AccessWe introduce SaltiNet, a deep neural network for scan-path prediction trained on 360-degree images. The model is based on a temporal-aware novel representation of saliency information named the saliency volume. The first ... -
Magnetic resonance imaging and machine learning make a valuable combined tool for the screening of preclinical AD
(2017-07)
Article
Restricted access - publisher's policy -
Hierarchical stack filtering: a bitplane-based algorithm for massively parallel processors
(2017-03-15)
Article
Open AccessWith the development of novel parallel architectures for image processing, the implementation of well-known image operators needs to be reformulated to take advantage of the so-called massive parallelism. In this work, we ... -
Cost-effective active learning for melanoma segmentation
(2017)
Conference lecture
Open AccessWe propose a novel Active Learning framework capable to train effectively a convolutional neural network for semantic segmentation of medical imaging, with a limited amount of training labeled data. Our contribution is a ... -
3D hierarchical optimization for multi-view depth map coding
(2017-11-28)
Article
Restricted access - publisher's policyDepth data has a widespread use since the popularity of high resolution 3D sensors. In multi-view sequences, depth information is used to supplement the color data of each view. This article proposes a joint encoding of ... -
Hierarchical object detection with deep reinforcement learning
(IOS Press, 2017-11-23)
Part of book or chapter of book
Restricted access - publisher's policyDeep learning and image processing are two areas of great interest to academics and industry professionals alike. The areas of application of these two disciplines range widely, encompassing fields such as medicine, robotics, ... -
Object retrieval with deep convolutional features
(IOS Press, 2017-11-23)
Part of book or chapter of book
Restricted access - publisher's policyDeep learning and image processing are two areas of great interest to academics and industry professionals alike. The areas of application of these two disciplines range widely, encompassing fields such as medicine, robotics, ...
RSS 1.0