• Title/Summary/Keyword: Pooling System

Search Result 78, Processing Time 0.033 seconds

Deep Learning System based on Morphological Neural Network (몰포러지 신경망 기반 딥러닝 시스템)

  • Choi, Jong-Ho
    • The Journal of Korea Institute of Information, Electronics, and Communication Technology
    • /
    • v.12 no.1
    • /
    • pp.92-98
    • /
    • 2019
  • In this paper, we propose a deep learning system based on morphological neural network(MNN). The deep learning layers are morphological operation layer, pooling layer, ReLU layer, and the fully connected layer. The operations used in morphological layer are erosion, dilation, and edge detection, etc. Unlike CNN, the number of hidden layers and kernels applied to each layer is limited in MNN. Because of the reduction of processing time and utility of VLSI chip design, it is possible to apply MNN to various mobile embedded systems. MNN performs the edge and shape detection operations with a limited number of kernels. Through experiments using database images, it is confirmed that MNN can be used as a deep learning system and its performance.

Deep Learning Architectures and Applications (딥러닝의 모형과 응용사례)

  • Ahn, SungMahn
    • Journal of Intelligence and Information Systems
    • /
    • v.22 no.2
    • /
    • pp.127-142
    • /
    • 2016
  • Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

Multiscale Spatial Position Coding under Locality Constraint for Action Recognition

  • Yang, Jiang-feng;Ma, Zheng;Xie, Mei
    • Journal of Electrical Engineering and Technology
    • /
    • v.10 no.4
    • /
    • pp.1851-1863
    • /
    • 2015
  • – In the paper, to handle the problem of traditional bag-of-features model ignoring the spatial relationship of local features in human action recognition, we proposed a Multiscale Spatial Position Coding under Locality Constraint method. Specifically, to describe this spatial relationship, we proposed a mixed feature combining motion feature and multi-spatial-scale configuration. To utilize temporal information between features, sub spatial-temporal-volumes are built. Next, the pooled features of sub-STVs are obtained via max-pooling method. In classification stage, the Locality-Constrained Group Sparse Representation is adopted to utilize the intrinsic group information of the sub-STV features. The experimental results on the KTH, Weizmann, and UCF sports datasets show that our action recognition system outperforms the classical local ST feature-based recognition systems published recently.

A Conceptual Framework for Interorganizational Systems Based on Linkage of Value Chain Activities (기업 가치활동의 연계에 근거한 조직간 정보시스템의 개념적 틀)

  • Hong, Il-Yoo;Kim, Chang-Su
    • Asia pacific journal of information systems
    • /
    • v.10 no.4
    • /
    • pp.21-36
    • /
    • 2000
  • This paper addresses the need to incorporate the increasing trend of partnership formation among business firms into a framework for interorganizational systems. The existing frameworks for classifying interorganizational systems are either conceptually too complex to be readily applicable to IOS planning or too outdated to explain the emergence of numerous forms of global communication networks. Based upon two dimensions, namely value activity linkage and system support level, we propose a new IOS framework which classifies IOS into four types, including (1) operational cooperation, (2) resource pooling, (3) operational coordination, and (4) complementary partnership. We review select cases that fit into each IOS category and draw characteristics of systems of each category. The paper concludes with suggestions for applying the framework to the development of an IOS-enabled corporate strategy.

  • PDF

Performance Improvement of Object Recognition System in Broadcast Media Using Hierarchical CNN (계층적 CNN을 이용한 방송 매체 내의 객체 인식 시스템 성능향상 방안)

  • Kwon, Myung-Kyu;Yang, Hyo-Sik
    • Journal of Digital Convergence
    • /
    • v.15 no.3
    • /
    • pp.201-209
    • /
    • 2017
  • This paper is a smartphone object recognition system using hierarchical convolutional neural network. The overall configuration is a method of communicating object information to the smartphone by matching the collected data by connecting the smartphone and the server and recognizing the object to the convergence neural network in the server. It is also compared to a hierarchical convolutional neural network and a fractional convolutional neural network. Hierarchical convolutional neural networks have 88% accuracy, fractional convolutional neural networks have 73% accuracy and 15%p performance improvement. Based on this, it shows possibility of expansion of T-Commerce market connected with smartphone and broadcasting media.

Development of Infrared-Ray Communication System for Position Recognition of Yard Tractor in Container Terminal (컨테이너터미널 내의 야드 트랙터 위치인식을 위한 적외선 통신시스템 개발)

  • Hong, Dong-Hee;Kim, Chang-Gon
    • Journal of Digital Convergence
    • /
    • v.11 no.1
    • /
    • pp.211-223
    • /
    • 2013
  • In Korea, the location of yard tractors is figured out in real time by using RFID system in container terminals. However, even though the location recognition of RFID system works fine when transfer crane is in yard operation, there are some problems when container crane is in ship operation. That is because yard tractors come one by one to each transfer crane in an order, but yard tractors come in 4 lanes to the container crane, which makes the system impossible to recognize each yard tractor separately. Therefore, we developed the infrared-ray communication system which can recognize yard tractors accurately in not only in the yard operation of transfer crane but also in the ship operation of container crane in same way in this study. The result in this study showed constant number of recognition, and the range of recognition measures 5.7m in 25m distance. The range of recognition shown in this study is enough to recognize each yard tractor passing under container crane separately.

Train Speed Control in Slope Area Using Infrared System (적외선 시스템을 이용한 경사 지역에서 열차 운행 속도 제어)

  • Sugiana, Ahmad;Sanyoto, Mulyo;Parwito, Parwito;Agrianto, Yanardian;Lee, Key Seo;Choy, Ick
    • The Journal of the Korea institute of electronic communication sciences
    • /
    • v.11 no.6
    • /
    • pp.635-644
    • /
    • 2016
  • Train speed control is a vital part of train protection to build safe movement at an operation track. There is a special condition of track that needs more attention to protect the train, for example in slope area. Moreover, in developing country with vandalism problem, it requires to install minimalized equipment on the trackside. In addition, in tropical country, on tracksides it will be potentially pooling water that influences to the performance of trackside equipment. To address these problems, we propose the train speed control for slope area using infrared system. By installing on the pole configuration, the system offers a less challenging, economically sensible, minimalized installation of equipment on the trackside and reliability for heavy rain environment. This paper concentrates on the controlling train speed and measurement performance evaluation in slope area. The proposed train speed control system can monitor and control the speed in sloping area with maximum 3.6% and controlled speed about 20 km per h.

CNN Based 2D and 2.5D Face Recognition For Home Security System (홈보안 시스템을 위한 CNN 기반 2D와 2.5D 얼굴 인식)

  • MaYing, MaYing;Kim, Kang-Chul
    • The Journal of the Korea institute of electronic communication sciences
    • /
    • v.14 no.6
    • /
    • pp.1207-1214
    • /
    • 2019
  • Technologies of the 4th industrial revolution have been unknowingly seeping into our lives. Many IoT based home security systems are using the convolutional neural network(CNN) as good biometrics to recognize a face and protect home and family from intruders since CNN has demonstrated its excellent ability in image recognition. In this paper, three layouts of CNN for 2D and 2.5D image of small dataset with various input image size and filter size are explored. The simulation results show that the layout of CNN with 50*50 input size of 2.5D image, 2 convolution and max pooling layer, and 3*3 filter size for small dataset of 2.5D image is optimal for a home security system with recognition accuracy of 0.966. In addition, the longest CPU time consumption for one input image is 0.057S. The proposed layout of CNN for a face recognition is suitable to control the actuators in the home security system because a home security system requires good face recognition and short recognition time.

The Teachers' Recognition and a Plan for the Improvement of the System on Selection of Gifted Students in Science Using Teachers' Observation and Nomination (과학 영재 관찰.추천 선발 방식에 대한 교사의 인식 조사 및 개선 방안)

  • Bang, Mi Seon;Kim, Yong Gwon
    • Journal of Korean Elementary Science Education
    • /
    • v.32 no.2
    • /
    • pp.169-184
    • /
    • 2013
  • The purpose of this study is to investigate teachers' recognition and to suggest an improvement in the system of teacher's observation and nomination used to selecting gifted and talented students in Science in the Busan Metropolitan School District in 2013 by investigating teachers' recognition of the system and their expressed needs. The results are as follows. First, it was observed that teachers are of the opinion that it is difficult to determine the science gifted students by observation due to their lack of expertise in giftedness and gifted education, the lack of a check list to use, and the difficulty of ensuring the objectivity of the results of the determination. Second, the absence of objective screening tools used for the selection, the selection of gifted students based on their subjective judgment, and the possibility to select students based only on visible manifestations of ability may cause parents to mistrust the system. Thus, institutional support is required to address the concerns of teachers and parents. Third, the teachers who are in charge of observation, nomination, selection and determination need to be trained. After that, at least one of these teachers should be assigned in each school and training should operate continuously and systematically. Lastly, while these things are occurring, the process of observation and nomination of by teachers, which is the basis of pooling gifted students at the level of Busan Metropolitan School District, should be continued.

Optimal Design of Generalized Process-storage Network Applicable To Polymer Processes (고분자 공정에 적용할 수 있는 일반화된 공정-저장조 망구조 최적설계)

  • Yi, Gyeongbeom;Lee, Euy-Soo
    • Korean Chemical Engineering Research
    • /
    • v.45 no.3
    • /
    • pp.249-257
    • /
    • 2007
  • The periodic square wave (PSW) model was successfully applied to the optimal design of a batch-storage network. The network structure can cover any type of batch production, distribution and inventory system, including recycle streams. Here we extend the coverage of the PSW model to multitasking semi-continuous processes as well as pure continuous and batch processes. In previous solutions obtained using the PSW model, the feedstock composition and product yield were treated as known constants. This constraint is relaxed in the present work, which treats the feedstock composition and product yield as free variables to be optimized. This modification makes it possible to deal with the pooling problem commonly encountered in oil refinery processes. Despite the greater complexity that arises when the feedstock composition and product yield are free variables, the PSW model still gives analytic lot sizing equations. The ability of the proposed method to determine the optimal plant design is demonstrated through the example of a high density polyethylene (HDPE) plant. Based on the analytical optimality results, we propose a practical process optimality measure that can be used for any kind of process. This measure facilitates direct comparison of the performance of multiple processes, and hence is a useful tool for diagnosing the status of process systems. The result that the cost of a process is proportional to the square root of average flow rate is similar to the well-known six-tenths factor rule in plant design.