Optimum Network/Framework Selection from High-Level Specifications in Embedded Deep Learning Vision Applications

2018 ◽  
pp. 369-379
Author(s):  
Delia Velasco-Montero ◽  
Jorge Fernández-Berni ◽  
Ricardo Carmona-Galán ◽  
Ángel Rodríguez-Vázquez
Keyword(s):  
Deep Learning ◽  
High Level

scholarly journals Deep-Framework: A Distributed, Scalable, and Edge-Oriented Framework for Real-Time Analysis of Video Streams

Sensors ◽  
2021 ◽  
Vol 21 (12) ◽  
pp. 4045
Author(s):  
Alessandro Sassu ◽  
Jose Francisco Saenz-Cogollo ◽  
Maurizio Agelli
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Video Data ◽  
Video Analytics ◽  
Web Based ◽  
Real Time Analysis ◽  
Open Source Framework ◽  
Cluster Configuration ◽  
Time Requirements ◽  
High Level

Edge computing is the best approach for meeting the exponential demand and the real-time requirements of many video analytics applications. Since most of the recent advances regarding the extraction of information from images and video rely on computation heavy deep learning algorithms, there is a growing need for solutions that allow the deployment and use of new models on scalable and flexible edge architectures. In this work, we present Deep-Framework, a novel open source framework for developing edge-oriented real-time video analytics applications based on deep learning. Deep-Framework has a scalable multi-stream architecture based on Docker and abstracts away from the user the complexity of cluster configuration, orchestration of services, and GPU resources allocation. It provides Python interfaces for integrating deep learning models developed with the most popular frameworks and also provides high-level APIs based on standard HTTP and WebRTC interfaces for consuming the extracted video data on clients running on browsers or any other web-based platform.

scholarly journals Building Extraction in Very High Resolution Imagery by Dense-Attention Networks

2018 ◽  
Vol 10 (11) ◽  
pp. 1768 ◽  
Author(s):  
Hui Yang ◽  
Penghai Wu ◽  
Xuedong Yao ◽  
Yanlan Wu ◽  
Biao Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Building Extraction ◽  
Learning Networks ◽  
Feature Maps ◽  
Low Level ◽  
High Resolution Imagery ◽  
Very High Resolution Imagery ◽  
High Level ◽  
Very High

Building extraction from very high resolution (VHR) imagery plays an important role in urban planning, disaster management, navigation, updating geographic databases, and several other geospatial applications. Compared with the traditional building extraction approaches, deep learning networks have recently shown outstanding performance in this task by using both high-level and low-level feature maps. However, it is difficult to utilize different level features rationally with the present deep learning networks. To tackle this problem, a novel network based on DenseNets and the attention mechanism was proposed, called the dense-attention network (DAN). The DAN contains an encoder part and a decoder part which are separately composed of lightweight DenseNets and a spatial attention fusion module. The proposed encoder–decoder architecture can strengthen feature propagation and effectively bring higher-level feature information to suppress the low-level feature and noises. Experimental results based on public international society for photogrammetry and remote sensing (ISPRS) datasets with only red–green–blue (RGB) images demonstrated that the proposed DAN achieved a higher score (96.16% overall accuracy (OA), 92.56% F1 score, 90.56% mean intersection over union (MIOU), less training and response time and higher-quality value) when compared with other deep learning methods.

scholarly journals A Survey of Graphical Page Object Detection with Deep Neural Networks

2021 ◽  
Author(s):  
Jwalin Bhatt ◽  
Khurram Azeem Hashmi ◽  
Muhammad Zeshan Afzal ◽  
Didier Stricker
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Conceptual Understanding ◽  
Deep Neural Networks ◽  
State Of The Art ◽  
Learning Approaches ◽  
Document Images ◽  
Essential Information ◽  
Current State ◽  
High Level

In any document, graphical elements like tables, figures, and formulas contain essential information. The processing and interpretation of such information require specialized algorithms. Off-the-shelf OCR components cannot process this information reliably. Therefore, an essential step in document analysis pipelines is to detect these graphical components. It leads to a high-level conceptual understanding of the documents that makes digitization of documents viable. Since the advent of deep learning, the performance of deep learning-based object detection has improved many folds. In this work, we outline and summarize the deep learning approaches for detecting graphical page objects in the document images. Therefore, we discuss the most relevant deep learning-based approaches and state-of-the-art graphical page object detection in document images. This work provides a comprehensive understanding of the current state-of-the-art and related challenges. Furthermore, we discuss leading datasets along with the quantitative evaluation. Moreover, it discusses briefly the promising directions that can be utilized for further improvements.

scholarly journals Deep Learning Based BCI Control of a Robotic Service Assistant Using Intelligent Goal Formulation

2018 ◽  
Author(s):  
D. Kuhner ◽  
L.D.J. Fiederer ◽  
J. Aldinger ◽  
F. Burget ◽  
M. Völker ◽  
...  
Keyword(s):  
Deep Learning ◽  
User Study ◽  
Human Robot Interaction ◽  
Service Robots ◽  
Referring Expressions ◽  
Non Invasive ◽  
Neuronal Signal ◽  
High Level ◽  
User Friendly ◽  
Robotic Tasks

AbstractAs autonomous service robots become more affordable and thus available for the general public, there is a growing need for user-friendly interfaces to control these systems. Control interfaces typically get more complicated with increasing complexity of the robotic tasks and the environment. Traditional control modalities as touch, speech or gesture commands are not necessarily suited for all users. While non-expert users can make the effort to familiarize themselves with a robotic system, paralyzed users may not be capable of controlling such systems even though they need robotic assistance most. In this paper, we present a novel framework, that allows these users to interact with a robotic service assistant in a closed-loop fashion, using only thoughts. The system is composed of several interacting components: non-invasive neuronal signal recording and co-adaptive deep learning which form the brain-computer interface (BCI), high-level task planning based on referring expressions, navigation and manipulation planning as well as environmental perception. We extensively evaluate the BCI in various tasks, determine the performance of the goal formulation user interface and investigate its intuitiveness in a user study. Furthermore, we demonstrate the applicability and robustness of the system in real world scenarios, considering fetch-and-carry tasks and tasks involving human-robot interaction. As our results show, the system is capable of adapting to frequent changes in the environment and reliably accomplishes given tasks within a reasonable amount of time. Combined with high-level planning using referring expressions and autonomous robotic systems, interesting new perspectives open up for non-invasive BCI-based human-robot interactions.

scholarly journals The Experience of Deep Learning by Accounting Students in a University Accounting Course

2021 ◽  
Author(s):  
◽  
Martin Craig Turner
Keyword(s):  
Deep Learning ◽  
Intrinsic Motivation ◽  
Creative Thinking ◽  
Cognitive Effort ◽  
Accounting Students ◽  
Teacher Student ◽  
Education Theory ◽  
Psychology Literature ◽  
Student Relationship ◽  
High Level

<p>Higher education in accounting faces a challenge to shift its emphasis from reproducing technical knowledge to developing personal capabilities such as critical thinking, creative thinking, problem-solving, communication and teamwork. The educational psychology literature suggests students will not make the cognitive effort to develop personal capabilities unless they frrst experience a deep approach to learning; and the experience of high-level relevance structure, high-level conception of learning and intrinsic motivation strongly support deep learning. This study examines how accounting students can be supported to experience high-level relevance structure, high-level conception of learning, intrinsic motivation and deep learning in the context of a university accounting course. Phenomenography is used to study the experience of learning of students in a third year undergraduate accounting course into which an integrated set of interventions involving Assessment, Teamwork, Teacher-Student Relationship and Instruction was introduced. An assignment in five stages and five session preparation assignments, supplemented with a focus group and surveys of students, identified and captured the ways students experience key aspects of how they learn. A key fmding of this study is that it is possible to transform the design and delivery of a single university course to support a large proportion of students to experience change in how they learn accounting (and, in particular, to experience deep learning) through the careful adaptation of education theory. An implication of this study is the need to support students to experience change in how they learn in frrst year courses to enable them to develop personal capabilities in their later university studies.</p>

scholarly journals Deep Learning Strategies for ProtoDUNE Raw Data Denoising

2022 ◽  
Vol 6 (1) ◽  
Author(s):  
Marco Rossi ◽  
Sofia Vallecorsa
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Learning Strategies ◽  
Simulation Data ◽  
Raw Data ◽  
Digital Detector ◽  
Speed Up ◽  
Neural Network Hardware ◽  
Graph Neural Networks ◽  
High Level

AbstractIn this work, we investigate different machine learning-based strategies for denoising raw simulation data from the ProtoDUNE experiment. The ProtoDUNE detector is hosted by CERN and it aims to test and calibrate the technologies for DUNE, a forthcoming experiment in neutrino physics. The reconstruction workchain consists of converting digital detector signals into physical high-level quantities. We address the first step in reconstruction, namely raw data denoising, leveraging deep learning algorithms. We design two architectures based on graph neural networks, aiming to enhance the receptive field of basic convolutional neural networks. We benchmark this approach against traditional algorithms implemented by the DUNE collaboration. We test the capabilities of graph neural network hardware accelerator setups to speed up training and inference processes.

Deep Learning Network

2020 ◽  
pp. 1-30
Author(s):  
Bhanu Chander
Keyword(s):  
Artificial Intelligence ◽  
Machine Learning ◽  
Deep Learning ◽  
Great Power ◽  
Human Being ◽  
Technology Improvement ◽  
Deep Learning Network ◽  
High Level ◽  
Human Brains ◽  
Near Future

Artificial intelligence (AI) is defined as a machine that can do everything a human being can do and produce better results. Means AI enlightening that data can produce a solution for its own results. Inside the AI ellipsoidal, Machine learning (ML) has a wide variety of algorithms produce more accurate results. As a result of technology, improvement increasing amounts of data are available. But with ML and AI, it is very difficult to extract such high-level, abstract features from raw data, moreover hard to know what feature should be extracted. Finally, we now have deep learning; these algorithms are modeled based on how human brains process the data. Deep learning is a particular kind of machine learning that provides flexibility and great power, with its attempts to learn in multiple levels of representation with the operations of multiple layers. Deep learning brief overview, platforms, Models, Autoencoders, CNN, RNN, and Appliances are described appropriately. Deep learning will have many more successes in the near future because it requires very little engineering by hand.

scholarly journals Natural Language Understanding for Multi-Level Distributed Intelligent Virtual Sensors

IoT ◽  
2020 ◽  
Vol 1 (2) ◽  
pp. 494-505
Author(s):  
Radu-Casian Mihailescu ◽  
Georgios Kyriakou ◽  
Angelos Papangelis
Keyword(s):  
Deep Learning ◽  
User Interface ◽  
Natural Language ◽  
Learning Approach ◽  
Language Understanding ◽  
Virtual Sensor ◽  
Virtual Sensors ◽  
Multi Level ◽  
High Level ◽  
User Friendly

In this paper we address the problem of automatic sensor composition for servicing human-interpretable high-level tasks. To this end, we introduce multi-level distributed intelligent virtual sensors (multi-level DIVS) as an overlay framework for a given mesh of physical and/or virtual sensors already deployed in the environment. The goal for multi-level DIVS is two-fold: (i) to provide a convenient way for the user to specify high-level sensing tasks; (ii) to construct the computational graph that provides the correct output given a specific sensing task. For (i) we resort to a conversational user interface, which is an intuitive and user-friendly manner in which the user can express the sensing problem, i.e., natural language queries, while for (ii) we propose a deep learning approach that establishes the correspondence between the natural language queries and their virtual sensor representation. Finally, we evaluate and demonstrate the feasibility of our approach in the context of a smart city setup.

scholarly journals Deep Learning-Based Classification of Weld Surface Defects

2019 ◽  
Vol 9 (16) ◽  
pp. 3312 ◽  
Author(s):  
Zhu ◽  
Ge ◽  
Liu
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Random Forest ◽  
Convolutional Neural Network ◽  
Surface Defects ◽  
Recognition Method ◽  
High Level ◽  
Intelligent Recognition ◽  
Non Destructive

In order to realize the non-destructive intelligent identification of weld surface defects, an intelligent recognition method based on deep learning is proposed, which is mainly formed by convolutional neural network (CNN) and forest random. First, the high-level features are automatically learned through the CNN. Random forest is trained with extracted high-level features to predict the classification results. Secondly, the weld surface defects images are collected and preprocessed by image enhancement and threshold segmentation. A database of weld surface defects is established using pre-processed images. Finally, comparative experiments are performed on the weld surface defects database. The results show that the accuracy of the method combined with CNN and random forest can reach 0.9875, and it also demonstrates the method is effective and practical.

Sign in / Sign up

Export Citation Format

Share Document