Suchergebnisse

Convolutional Neural Networks (CNNs) have shown to yield very strong results in several Computer Vision tasks. Their application to language has received much less attention, and it has mainly focused on static classification tasks, such as sentence classification for Sentiment Analysis or relation extraction. In this work, we study the application of CNNs to language modeling, a dynamic, sequential prediction task that needs models to capture local as well as long-range dependency information. Our contribution is twofold. First, we show that CNNs achieve 11-26% better absolute performance than feed-forward neural language models, demonstrating their potential for language representation even in sequential tasks. As for recurrent models, our model outperforms RNNs but is below state of the art LSTM models. Second, we gain some understanding of the behavior of the model, showing that CNNs in language act as feature detectors at a high level of abstraction, like in Computer Vision, and that the model can profitably use information from as far as 16 words before the target. ; This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 655577 (LOVe); ERC 2011 Starting Independent Research Grant n. 283554 (COMPOSES) and the Erasmus Mundus Scholarship for Joint Master Programs.

Comunicació presentada al CVPR 2019: Conference on Computer Vision and Pattern Recognition, celebrat del 16 al 20 de juny de 2019 a California, Estats Units d'Amèrica. ; Visual illusions teach us that what we see is not always what is represented in the physical world. Their special nature make them a fascinating tool to test and validate any new vision model proposed. In general, current vision models are based on the concatenation of linear and non-linear operations. The similarity of this structure with the operations present in Convolutional Neural Networks (CNNs) has motivated us to study if CNNs trained for low-level visual tasks are deceived by visual illusions. In particular, we show that CNNs trained for image denoising, image deblurring, and computational color constancy are able to replicate the human response to visual illusions, and that the extent of this replication varies with respect to variation in architecture and spatial pattern size. These results suggest that in order to obtain CNNs that better replicate human behaviour, we may need to start aiming for them to better replicate visual illusions. ; This work has received funding from the EU Horizon 2020 programme under grant agreement 761544 (project HDR4EU) and under grant agreement 780470 (project SAUCE) and by the Spanish government and FEDER Fund, grant ref. TIN2015-71537-P (MINECO/FEDER,UE). We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research.

The classification of logos is a particular case within computer vision since they have their own characteristics. Logos can contain only text, iconic images or a combination of both, and they usually include figurative symbols designed by experts that vary substantially besides they may share the same semantics. This work presents a method for multi-label classification and retrieval of logo images. For this, Convolutional Neural Networks (CNN) are trained to classify logos from the European Union TradeMark (EUTM) dataset according to their colors, shapes, sectors and figurative designs. An auto-encoder is also trained to learn representations of the input images. Once trained, the neural codes from the last convolutional layers in the CNN and the central layer of the auto-encoder can be used to perform similarity search through kNN, allowing us to obtain the most similar logos based on their color, shape, sector, figurative elements, overall features, or a weighted combination of them provided by the user. To the best of our knowledge, this is the first multi-label classification method for logos, and the only one that allows retrieving a ranking of images with these criteria provided by the user. ; This work is supported by the Spanish Ministry HISPAMUS project with code TIN2017-86576-R, partially funded by the EU.

Pedestrian movement direction recognition is an important factor in autonomous driver assistance and security surveillance systems. Pedestrians are the most crucial and fragile moving objects in streets, roads, and events, where thousands of people may gather on a regular basis. People flow analysis on zebra crossings and in shopping centers or events such as demonstrations are a key element to improve safety and to enable autonomous cars to drive in real life environments. This paper focuses on deep learning techniques such as convolutional neural networks (CNN) to achieve a reliable detection of pedestrians moving in a particular direction. We propose a CNN-based technique that leverages current pedestrian detection techniques (histograms of oriented gradients-linSVM) to generate a sum of subtracted frames (flow estimation around the detected pedestrian), which are used as an input for the proposed modified versions of various state-of-the-art CNN networks, such as AlexNet, GoogleNet, and ResNet. Moreover, we have also created a new data set for this purpose, and analyzed the importance of training in a known data set for the neural networks to achieve reliable results. ; This work was supported by the Feder funds, Spanish Government through the COMBAHO Project, under Grant TIN2016-76515-R, and in part by the University of Alicante Project under Grant GRE16-19.

Phytoplankton form the basis of the marine food web and are an indicator for the overall status of the marine ecosystem. Changes in this community may impact a wide range of species (Capuzzo et al. 2018) ranging from zooplankton and fish to seabirds and marine mammals. Efficient monitoring of the phytoplankton community is therefore essential (Edwards et al. 2002). Traditional monitoring techniques are highly time intensive and involve taxonomists identifying and counting numerous specimens under the light microscope. With the recent development of automated sampling devices, image analysis technologies and learning algorithms, the rate of counting and identification of phytoplankton can be increased significantly (Thyssen et al. 2015). The FlowCAM (Álvarez et al. 2013) is an imaging particle analysis system for the identification and classification of phytoplankton. Within the Belgian Lifewatch observatory, monthly phytoplankton samples are taken at nine stations in the Belgian part of the North Sea. These samples are run through the FlowCAM and each particle is photographed. Next, the particles are identified based on their morphology (and fluorescence) using state-of-the-art Convolutional Neural Networks (CNNs) for computer vision. This procedure requires learning sets of expert validated images. The CNNs are specifically designed to take advantage of the two dimensional structure of these images by finding local patterns, being easier to train and having many fewer parameters than a fully connected network with the same number of hidden units. In this work we present our approach to the use of CNNs for the identification and classification of phytoplankton, testing it on several benchmarks and comparing with previous classification techniques. The network architecture used is ResNet50 (He et al. 2016). The framework is fully written in Python using the TensorFlow (Abadi, M. et al. 2016) module for Deep Learning. Deployment and exploitation of the current framework is supported by the recently started European Union Horizon 2020 programme funded project DEEP-Hybrid-Datacloud (Grant Agreement number 777435), which supports the expensive training of the system needed to develop the application and provides the necessary computational resources to the users.

One of the riches of the Indonesian nation comes from the diversity of ethnicities and cultures, especially dance, which is the culture of the Indonesian people, starting from their ancestors until now, their authenticity is still maintained. The wrong cultural dance that develops, especially in South Sulawesi, which consists of four (4) ethnic groups, namely: Bugis, Makassar, Toraja and Mandar, which have their own dance dances from each tribe in South Sulawesi to maintain this dance. There is a need for collaboration between local community leaders, government and researchers, especially researchers to raise dance dances from the Makassar Tribe called Pakkarena dance using the Convolutional Neural Network (CNN) method to the Pakarena dance image in distinguishing or classifying an object on digital images with an accuracy level of 95 75%.

Convolutional Neural Networks (CNN) are the most popular of deep network models due to their applicability and success in image processing. Although plenty of effort has been made in designing and training better discriminative CNNs, little is yet known about the internal features these models learn. Questions like, what specific knowledge is coded within CNN layers, and how can it be used for other purposes besides discrimination, remain to be answered. To advance in the resolution of these questions, in this work we extract features from CNN layers, building vector representations from CNN activations. The resultant vector embedding is used to represent first images and then known image classes. On those representations we perform an unsupervised clustering process, with the goal of studying the hidden semantics captured in the embedding space. Several abstract entities untaught to the network emerge in this process, effectively defining a taxonomy of knowledge as perceived by the CNN. We evaluate and interpret these sets using WordNet, while studying the different behaviours exhibited by the layers of a CNN model according to their depth. Our results indicate that, while top (i.e., deeper) layers provide the most representative space, low layers also define descriptive dimensions. ; This work was partially supported by the IBM/BSC Technology Center for Supercomputing (Joint Study Agreement, No. W156463), by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051). ; Peer Reviewed ; Postprint (author's final draft)

Convolutional Neural Networks (CNN) are the most popular of deep network models due to their applicability and success in image processing. Although plenty of effort has been made in designing and training better discriminative CNNs, little is yet known about the internal features these models learn. Questions like, what specific knowledge is coded within CNN layers, and how can it be used for other purposes besides discrimination, remain to be answered. To advance in the resolution of these questions, in this work we extract features from CNN layers, building vector representations from CNN activations. The resultant vector embedding is used to represent first images and then known image classes. On those representations we perform an unsupervised clustering process, with the goal of studying the hidden semantics captured in the embedding space. Several abstract entities untaught to the network emerge in this process, effectively defining a taxonomy of knowledge as perceived by the CNN. We evaluate and interpret these sets using WordNet, while studying the different behaviours exhibited by the layers of a CNN model according to their depth. Our results indicate that, while top (i.e., deeper) layers provide the most representative space, low layers also define descriptive dimensions. ; This work was partially supported by the IBM/BSC Technology Center for Supercomputing (Joint Study Agreement, No. W156463), by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology through TIN2015-65316-P project and by the Generalitat de Catalunya (contracts 2014-SGR-1051). ; Peer Reviewed ; Postprint (author's final draft)

Ships are an integral part of maritime traffic where they play both militaries as well as non-combatant roles. This vast maritime traffic needs to be managed and monitored by identifying and recognising vessels to ensure the maritime safety and security. As an approach to find an automated and efficient solution, a deep learning model exploiting convolutional neural network (CNN) as a basic building block, has been proposed in this paper. CNN has been predominantly used in image recognition due to its automatic high-level features extraction capabilities and exceptional performance. We have used transfer learning approach using pre-trained CNNs based on VGG16 architecture to develop an algorithm that performs the different ship types classification. This paper adopts data augmentation and fine-tuning to further improve and optimize the baseline VGG16 model. The proposed model attains an average classification accuracy of 97.08% compared to the average classification accuracy of 88.54% obtained from the baseline model.

Phytoplankton form the basis of the marine food web and are an indicator for the overall status of the marine ecosystem. Changes in this community may impact a wide range of species (Capuzzo et al. 2018) ranging from zooplankton and fish to seabirds and marine mammals. Efficient monitoring of the phytoplankton community is therefore essential (Edwards et al. 2002). Traditional monitoring techniques are highly time intensive and involve taxonomists identifying and counting numerous specimens under the light microscope. With the recent development of automated sampling devices, image analysis technologies and learning algorithms, the rate of counting and identification of phytoplankton can be increased significantly (Thyssen et al. 2015). The FlowCAM (Álvarez et al. 2013) is an imaging particle analysis system for the identification and classification of phytoplankton. Within the Belgian Lifewatch observatory, monthly phytoplankton samples are taken at nine stations in the Belgian part of the North Sea. These samples are run through the FlowCAM and each particle is photographed. Next, the particles are identified based on their morphology (and fluorescence) using state-of-the-art Convolutional Neural Networks (CNNs) for computer vision. This procedure requires learning sets of expert validated images. The CNNs are specifically designed to take advantage of the two dimensional structure of these images by finding local patterns, being easier to train and having many fewer parameters than a fully connected network with the same number of hidden units.In this work we present our approach to the use of CNNs for the identification and classification of phytoplankton, testing it on several benchmarks and comparing with previous classification techniques. The network architecture used is ResNet50 (He et al. 2016). The framework is fully written in Python using the TensorFlow (Abadi, M. et al. 2016) module for Deep Learning.Deployment and exploitation of the current framework is supported by the recently started European Union Horizon 2020 programme funded project DEEP-Hybrid-Datacloud (Grant Agreement number 777435), which supports the expensive training of the system needed to develop the application and provides the necessary computational resources to the users.