The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Analitis video biasanya memakan masa kerana ia bukan sahaja memerlukan penyahkodan video sebagai langkah pertama tetapi juga biasanya menggunakan penglihatan komputer yang kompleks dan algoritma pembelajaran mesin pada bingkai yang dinyahkodkan. Untuk mencapai kecekapan tinggi dalam analisis video dengan saiz bingkai yang semakin meningkat, banyak penyelidikan telah dijalankan untuk pemprosesan video yang diedarkan menggunakan Hadoop. Walau bagaimanapun, kebanyakan pendekatan tertumpu pada pemprosesan berbilang fail video pada berbilang nod. Pendekatan sedemikian memerlukan beberapa fail video untuk mencapai apa-apa kelajuan, dan boleh mengakibatkan ketidakseimbangan beban dengan mudah apabila saiz fail video agak panjang kerana fail video itu sendiri diproses secara berurutan. Sebaliknya, kami mencadangkan kaedah penyahkodan video yang diedarkan dengan FFmpeg lanjutan dan VideoRecordReader, yang dengannya satu fail video besar boleh diproses secara selari merentas berbilang nod dalam Hadoop. Keputusan eksperimen menunjukkan bahawa kajian kes pengesanan muka dan sistem SURF masing-masing mencapai 40.6 kali dan 29.1 kali kelajuan pada kelompok empat nod dengan 12 pemeta dalam setiap nod, menunjukkan kebolehskalaan yang baik.
Illo YOON
University of Seoul
Saehanseul YI
University of Seoul
Chanyoung OH
University of Seoul
Hyeonjin JUNG
University of Seoul
Youngmin YI
University of Seoul
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Salinan
Illo YOON, Saehanseul YI, Chanyoung OH, Hyeonjin JUNG, Youngmin YI, "Distributed Video Decoding on Hadoop" in IEICE TRANSACTIONS on Information,
vol. E101-D, no. 12, pp. 2933-2941, December 2018, doi: 10.1587/transinf.2018PAP0014.
Abstract: Video analytics is usually time-consuming as it not only requires video decoding as a first step but also usually applies complex computer vision and machine learning algorithms to the decoded frame. To achieve high efficiency in video analytics with ever increasing frame size, many researches have been conducted for distributed video processing using Hadoop. However, most approaches focused on processing multiple video files on multiple nodes. Such approaches require a number of video files to achieve any speedup, and could easily result in load imbalance when the size of video files is reasonably long since a video file itself is processed sequentially. In contrast, we propose a distributed video decoding method with an extended FFmpeg and VideoRecordReader, by which a single large video file can be processed in parallel across multiple nodes in Hadoop. The experimental results show that a case study of face detection and SURF system achieve 40.6 times and 29.1 times of speedups respectively on a four-node cluster with 12 mappers in each node, showing good scalability.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018PAP0014/_p
Salinan
@ARTICLE{e101-d_12_2933,
author={Illo YOON, Saehanseul YI, Chanyoung OH, Hyeonjin JUNG, Youngmin YI, },
journal={IEICE TRANSACTIONS on Information},
title={Distributed Video Decoding on Hadoop},
year={2018},
volume={E101-D},
number={12},
pages={2933-2941},
abstract={Video analytics is usually time-consuming as it not only requires video decoding as a first step but also usually applies complex computer vision and machine learning algorithms to the decoded frame. To achieve high efficiency in video analytics with ever increasing frame size, many researches have been conducted for distributed video processing using Hadoop. However, most approaches focused on processing multiple video files on multiple nodes. Such approaches require a number of video files to achieve any speedup, and could easily result in load imbalance when the size of video files is reasonably long since a video file itself is processed sequentially. In contrast, we propose a distributed video decoding method with an extended FFmpeg and VideoRecordReader, by which a single large video file can be processed in parallel across multiple nodes in Hadoop. The experimental results show that a case study of face detection and SURF system achieve 40.6 times and 29.1 times of speedups respectively on a four-node cluster with 12 mappers in each node, showing good scalability.},
keywords={},
doi={10.1587/transinf.2018PAP0014},
ISSN={1745-1361},
month={December},}
Salinan
TY - JOUR
TI - Distributed Video Decoding on Hadoop
T2 - IEICE TRANSACTIONS on Information
SP - 2933
EP - 2941
AU - Illo YOON
AU - Saehanseul YI
AU - Chanyoung OH
AU - Hyeonjin JUNG
AU - Youngmin YI
PY - 2018
DO - 10.1587/transinf.2018PAP0014
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E101-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2018
AB - Video analytics is usually time-consuming as it not only requires video decoding as a first step but also usually applies complex computer vision and machine learning algorithms to the decoded frame. To achieve high efficiency in video analytics with ever increasing frame size, many researches have been conducted for distributed video processing using Hadoop. However, most approaches focused on processing multiple video files on multiple nodes. Such approaches require a number of video files to achieve any speedup, and could easily result in load imbalance when the size of video files is reasonably long since a video file itself is processed sequentially. In contrast, we propose a distributed video decoding method with an extended FFmpeg and VideoRecordReader, by which a single large video file can be processed in parallel across multiple nodes in Hadoop. The experimental results show that a case study of face detection and SURF system achieve 40.6 times and 29.1 times of speedups respectively on a four-node cluster with 12 mappers in each node, showing good scalability.
ER -