Akm Ashiquzzaman, Seungmin Oh, Dongsu Lee, Jihoon Lee, Jinsul Kim
{"title":"Compact Deeplearning Convolutional Neural Network based Hand Gesture Classifier Application for Smart Mobile Edge Computing","authors":"Akm Ashiquzzaman, Seungmin Oh, Dongsu Lee, Jihoon Lee, Jinsul Kim","doi":"10.1109/ICAIIC48513.2020.9065257","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065257","url":null,"abstract":"Modern Deeplearning Convolutional Neural Network-based image classifier are now becoming extremely high performing with acute accuracy. Although this provides a reliable classification system the sheer computational power for this type of classification makes it very difficult to run in remote environments lower computational power. In smart next- generation mobile edge computing platforms, the applications need to be lightweight and very robust and faster to provide classification results. In the context of the hand-operated smart factory, the classification of a given hand gesture is vital for its accurate operation. In this research, we have designed a hand gesture image collection and label model and later train a deeplearning convolutional neural network to classify this model with the highest 98% accuracy. Later we introduced the node pruning system improvement which compresses the model while retraining the same accuracy rate of 98%. This type of low weight pruned deeplearning convolutional neural network will help build a lightweight Mobile edge-based server with high accuracy and lower computation overheard based application that is suitable for smart factory based remote applications.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116961437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Cognitive Intelligence with Adaptive MAC Vehicular Ad-Hoc Network Optimization","authors":"J. Lim, Yoong Choon Chang, Dhaval K. Patel","doi":"10.1109/ICAIIC48513.2020.9065283","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065283","url":null,"abstract":"Successful and timely data delivery in evolutionary Vehicular Ad-Hoc Network (VANET) for Intelligent Transportation System (ITS) are particularly challenging due to the unique characteristics of VANET, such as fast topology change and frequent disruptions. The current VANET schemes use static parameters in an ever changing VANET topology, which results in lack of adaptability. Due to the fast topology change and frequent disruptions in VANET, a cognitive and adaptive VANET Medium Access Control (MAC) scheduling optimization protocol scheme, which is able to change its transmission parameters, to suit different topology change in VANET is preferable. In the proposed Adaptive V ANET MAC, namely AdMAC, change of signal to noise and interference ratio (SINR) is proposed as an indicator to represent VANET transmission environment and traffic density. The change of SINR is then used to determine the V ANET MAC contention window size. Simulation is conducted under congested and non-congested traffics, in the urban map of Kuala Lumpur, Malaysia. Results show that the proposed AdMAC has better adaptability, in terms of packet success rate and average delay.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126647190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
So Hasegawa, Ryoma Kitagawa, Takumi Ito, T. Nakajima, Song-Ju Kim, Y. Shoji, M. Hasegawa
{"title":"Performance Evaluation of Machine Learning Based Channel Selection Algorithm Implemented on IoT Sensor Devices and Its Application to Wireless Sensor Network for Building Monitoring System","authors":"So Hasegawa, Ryoma Kitagawa, Takumi Ito, T. Nakajima, Song-Ju Kim, Y. Shoji, M. Hasegawa","doi":"10.1109/ICAIIC48513.2020.9065211","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065211","url":null,"abstract":"The number of IoT devices has dramatically increased. Numerous IoT devices generate enormous traffic, which causes network congestions and packet losses. To deal with network congestions, we have proposed a channel selection algorithm based on reinforcement learning for IoT devices. We have implemented the algorithm on an IoT device with limited function and confirmed autonomous appropriate channel selection in a local area where devices are crowded. In this paper, we conduct evaluation experiments in real environment where devices are deployed distributedly and there are various IoT systems such as Sigfox, LoRaWAN competing for frequency bands. Our experimental results show that the proposed algorithm based on reinforcement learning improves packet delivery rate (frame success rates) and fairness of the network. Furthermore, we apply the machine learning based IoT devices to IoT sensor network for building monitoring. We deploy 30 sensor devices in a 4-floor building and confirm data collection with avoiding congestions and interference by proposed devices.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126913301","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Tung Lam Pham, Huy Nguyen, Hoan Nguyen, V. Nguyen, Van Bui, Y. Jang
{"title":"Object Detection Framework for High Mobility Vehicles Tracking in Night-Time","authors":"Tung Lam Pham, Huy Nguyen, Hoan Nguyen, V. Nguyen, Van Bui, Y. Jang","doi":"10.1109/ICAIIC48513.2020.9065060","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065060","url":null,"abstract":"Optical Vehicular Communication is one of the most promising areas in which Optical Camera Communication (OCC) involves as the core technology. Due to a lot of existing challenges in Optical Vehicular Communication, there is still a big gap between research works and commercialized products. Recent developments in computer vision and advanced image processing techniques could enhance the performance of vehicular OCC system significantly. Among those techniques, there are many of robust image processing tools which are developed based on the theory of Machine Learning, Deep Learning and Artificial Intelligence (AI), which could be applied in a wide range of applications including vehicular OCC application. In this paper, we would like to introduce our recent work on applying a well-known object detection framework - You Only Look Once (YOLO) to detect and track a large number of high mobility vehicles, which is also considered as region of interests (RoIs) in vehicular OCC system. From that, we conducted a test on the Korea highway at a raining night to analyze the effectiveness of this method in the vehicular OCC system.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115256995","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Survey of Missing Data Imputation Using Generative Adversarial Networks","authors":"Jaeyoon Kim, Donghyun Tae, Junhee Seok","doi":"10.1109/ICAIIC48513.2020.9065044","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065044","url":null,"abstract":"Recently, many deep learning models for missing data imputation have been studied. One of the most popular models is Generative Adversarial Networks (GANs), which generate plausible fake data through adversarial training. In this paper, we take a look at the architecture, objective of a generator and a discriminator, training method and loss function. After that, we can see what improvements have been made to each model. Moreover, we can easily compare several GAN-based models for missing data imputation.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114243203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Waveform-based End-to-end Deep Convolutional Neural Network with Multi-scale Sliding Windows for Weakly Labeled Sound Event Detection","authors":"Seokjin Lee, Minhan Kim","doi":"10.1109/ICAIIC48513.2020.9064985","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9064985","url":null,"abstract":"In this paper, a waveform-based end-to-end sound event detection algorithm that detects and classifies sound events using a deep convolutional neural network architecture is proposed. While most machine-learning-based acoustic signal processing systems utilize hand-crafted feature vectors e.g. log-Mel spectrogram, end-to-end methods, which utilize raw input data, have recently been investigated for use in various applications. Therefore, we develop an end-to-end architecture for sound event detection tasks with convolutional neural networks. The proposed model consists of multi-scale time frames and networks that handle both short and long signal characteristics; the frame slides by 0.1 second to provide a sufficiently fine resolution. The element network for each time frame consists of several one-dimensional convolutional neural networks with a deeply stacked structure. The results of the element networks are averaged and gated by sound activity detection. In order to handle unlabeled data, the trained networks are enhanced using the mean-teacher model. A decision is made via double thresholding, and the results are enhanced using class-wise minimum gap/length compensation. To evaluate our proposed approach, simulations are performed with development data from DCASE 2019 Task 4, and the results show that the proposed algorithm had a macro-averaged F1 score of 31.7% for the DCASE 2019 development dataset, 30.2% for the DCASE 2018 evaluation dataset, and 26.7% for the DCASE 2019 evaluation dataset.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122017410","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Tooth Detection and Segmentation with Mask R-CNN","authors":"Guohua Zhu, Zewen Piao, S. Kim","doi":"10.1109/ICAIIC48513.2020.9065216","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065216","url":null,"abstract":"In this paper, we expound the application of Mask RCNN on automatic tooth detection and segmentation. Mask R-CNN is a recently proposed surprising algorithm for object detection and semantic segmentation. This paper aims at detecting and segmenting tooth only. We show that Mask RCNN also has a good segmentation effect in complex and crowded teeth structures. We use the pixel accuracy (PA) to evaluate the results.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129036534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Automotive Ethernet-Based Black Box System","authors":"LU Yan-Xun, Yung-Yuan Chen, Wen-Nan Fang","doi":"10.1109/ICAIIC48513.2020.9065251","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065251","url":null,"abstract":"This paper proposes the idea of automotive Ethernet-based black box that can be used in intelligent vehicles to record the valuable information. We built the black box model on Matlab/Simulink which is tested by PreScan simulation platform to record vehicle's data, video and the other information during the simulation process. Then, we constructed the developed black box model using the physical heterogeneous network system and replaced PreScan camera model with automotive Ethernet AVB (Audio Video Bridging) camera to record the actual video via Ethernet. Finally, we use the experimental results of simulation platform and the physical platform to demonstrate the feasibility of the proposed idea.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"93 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129151652","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
S. Sunaga, K. Kikuma, Koki Jimbo, Koki Sato, K. Ueda
{"title":"A Method to Determine if a Test Case is Necessary using Support Vector Machine","authors":"S. Sunaga, K. Kikuma, Koki Jimbo, Koki Sato, K. Ueda","doi":"10.1109/ICAIIC48513.2020.9065275","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065275","url":null,"abstract":"Communication software used for Next Generation Networks (NGN) is required to have high reliability, so it incorporates many methods to improve quality and to prevent service interruptions during operation for public networks. However, the accumulation of quality improvement methods has resulted in problems with longer development times and increasing costs. The authors are working to solve such problems by automating development processes. In this paper, we propose a method to determine if a test case is necessary using support vector machine. We vectorize know-how regarding creation of verification test cases from experts who understand the requirement specifications. The support vector machine learns the vector as the training data. Then this is used to automate determination if a test case is necessary. We evaluated the effectiveness of the method through practical experiments, automatical determination if a test case of particular requirements is necessary.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130396753","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Measures of Complexity for Large Scale Image Datasets","authors":"A. Rahane, A. Subramanian","doi":"10.1109/ICAIIC48513.2020.9065274","DOIUrl":"https://doi.org/10.1109/ICAIIC48513.2020.9065274","url":null,"abstract":"Large scale image datasets are a growing trend in the field of machine learning. However, it is hard to quantitatively understand or specify how various datasets compare to each other - i.e., if one dataset is more complex or harder to “learn” with respect to a deep-learning based network. In this work, we build a series of relatively computationally simple methods to measure the complexity of a dataset. Furthermore, we present an approach to demonstrate visualizations of high dimensional data, in order to assist with visual comparison of datasets. We present our analysis using four datasets from the autonomous driving research community - Cityscapes, IDD, BDD and Vistas. Using entropy based metrics, we present a rank-order complexity of these datasets, which we compare with an established rank-order with respect to deep learning.","PeriodicalId":424681,"journal":{"name":"2020 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124339904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}