The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Our goal is to design architectures that retain the groundbreaking performance of CNNs for landmark localization and at the same time are lightweight, compact and suitable for applications with limited computational resources. To this end, we make the following contributions: (a) we are the first to study the effect of neural network binarization on localization tasks, namely human pose estimation...
Most of the prior works summarize videos by either exploring different heuristically designed criteria in an unsupervised way or developing fully supervised algorithms by leveraging human-crafted training data in form of video-summary pairs or importance annotations. However, unsupervised methods are blind to the video category and often fail to produce semantically meaningful video summaries. On...
Artificial neural networks and deep learning methodologies have had growing interest across industry domains, including IoT and mobile systems. However, in low-power applications, resource limitations and operating environment restrictions make implementations difficult. This survey examines efforts that target the data and compute challenges of implementing energy efficient, low cost, and accurate...
Several variants of the long short-term memory (LSTM) architecture for recurrent neural networks have been proposed since its inception in 1995. In recent years, these networks have become the state-of-the-art models for a variety of machine learning problems. This has led to a renewed interest in understanding the role and utility of various computational components of typical LSTM variants. In this...
Autoimmune diseases are the third cause of mortality in the world. The identification of anti-nuclear antibody (ANA) via Immunofluorescence (IIF) test in human epithelial type-2 cells (HEp-2) is a conventional method to support the diagnosis of such diseases. In the present work, three popular Convolutional Neural Networks (CNNs) are evaluated for this task: LeNet-5, AlexNet, and GoogLeNet. We also...
The visual and automatic classification of vehicles plays an important role in the Transport Area. Besides of security issues, the monitoring of the type of traffic in streets and highways, as well the traffic dynamics over time, allows the optimization of use and of resources related to such public infrastructure. In this work we propose a novel method, called 2D-DBM, for robust and efficient automatic...
Image dehazing can be described as the problem of mapping from a hazy image to a haze-free image. Most approaches to this problem use physical models based on simplifications and priors. In this work we demonstrate that a convolutional neural network with a deep architecture and a large image database is able to learn the entire process of dehazing, without the need to adjust parameters, resulting...
VGG 16 and Inception-v3 networks were trained using a texture dataset of muddied and clean cows. A new dataset with 600 images that is similar to the actual texture dataset was introduced and used to train the networks. The method used to train the networks was transfer learning. ImageNet weights were trained using the similar dataset, then the newly trained weights were trained again using the actual...
In the person re-identification across multiple camera research field, attributes of the pedestrian are important cues to differentiate the appearance of each identity. In this work, ten types of attributes are considered as defined in the DukeMTMC-attribute dataset. A custom deep network architecture is proposed to perform the identification process. Furthermore, experiments were carried out to assess...
In this paper, we propose a modified architecture of a Pi-Sigma Neural Network (PSNN) based on two modifications: extension of the activation function and adding delays to neurons in the hidden layer. These new networks are called respectively Activation Function Extended Pi-Sigma (AFEPS) and Delayed Pi-Sigma (DPS) are obtained first by adding an activation function to all hidden neurons and secondly...
We propose associative domain adaptation, a novel technique for end-to-end domain adaptation with neural networks, the task of inferring class labels for an unlabeled target domain based on the statistical properties of a labeled source domain. Our training scheme follows the paradigm that in order to effectively derive class labels for the target domain, a network should produce statistically domain...
For large-scale visual search, highly compressed yet meaningful representations of images are essential. Structured vector quantizers based on product quantization and its variants are usually employed to achieve such compression while minimizing the loss of accuracy. Yet, unlike binary hashing schemes, these unsupervised methods have not yet benefited from the supervision, end-to-end learning and...
We present an approach to accelerating a wide variety of image processing operators. Our approach uses a fully-convolutional network that is trained on input-output pairs that demonstrate the operator’s action. After training, the original operator need not be run at all. The trained network operates at full resolution and runs in constant time. We investigate the effect of network architecture on...
This paper proposed an innovative education platform-VREX (Virtual Reality based Education eXpansion), with combination of online and offline, to improve the curriculum building and teaching experience. VREX is based on Virtual Reality (VR) and we believe VR can revolutionize the education ecosystem. With some trials, we found VR can be used to promote curriculum effectiveness in an immersive environment...
Training of Artificial Neural Networks (ANN) is an important step to make the network able to accomplish the desired task. This capacity of learning in such networks makes them applied in many applications as modeling and control. However, many of training algorithms have some drawbacks like: too many parameters to be estimated, important calculus time. In this paper, we propose a very simple method...
This paper deals with the problem of audio source separation. To handle the complex and ill-posed nature of the problems of audio source separation, the current state-of-the-art approaches employ deep neural networks to obtain instrumental spectra from a mixture. In this study, we propose a novel network architecture that extends the recently developed densely connected convolutional network (DenseNet),...
In this paper, we use Diagonal Recurrent Neural Networks on a sequence prediction task. The modification from standard RNN is simple: Diagonal recurrent matrices are used instead of full. This results in better test likelihood and faster convergence compared to regular full RNNs in most of our experiments. We show the benefits of using diagonal recurrent matrices with popularly used LSTM and GRU architectures...
Deep learning (deep structured learning, hierarchical learning or deep machine learning) is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data by using multiple processing layers with complex structures or otherwise composed of multiple non-linear transformations. In this paper, we present the results of testing neural networks architectures...
In the realm of surface electromyography (sEMG) gesture recognition, deep learning algorithms are seldom employed. This is due in part to the large quantity of data required for them to train on. Consequently, it would be prohibitively time consuming for a single user to generate a sufficient amount of data for training such algorithms. In this paper, two datasets of 18 and 17 able-bodied participants...
Pedestrian detection is an important topic in object detection. Compared with other object detectors, YOLOv2 achieves high accuracy and fast speed for general object detection, however it degrades accuracy when detecting crowed pedestrians. In this paper, combining with the skip structure of FCN, we tailor the YOLOv2 network to improve the accuracy in detecting small pedestrians which appear in groups...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.