Recently, Super-resolution convolutional neural network (SRCNN) is widely known as a state of the art method for achieving single-image super resolution. However, performance problems such as jaggy and ringing artifacts exist in SRCNN. Moreover, in order to realize a real-time upconverting system for high-resolution video streams such as 4K/8K 60 fps, problems such as processing delay and implementation cost remain. In the present paper, we propose high-performance super-resolution via patch-based deep neural network (SR-PDNN) rather than a convolutional neural network (CNN). Despite the very simple end-to-end learning system, the SR-PDNN achieves higher performance than the conventional CNN-based approach. In addition, this system is suitable for ultra-low-delay video processing by hardware implementation using an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).
Reo AOKI
Kanazawa University,Visual Technologies (ASIC)
Kousuke IMAMURA
Kanazawa University
Akihiro HIRANO
Kanazawa University
Yoshio MATSUDA
Kanazawa University
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.
Copy
Reo AOKI, Kousuke IMAMURA, Akihiro HIRANO, Yoshio MATSUDA, "High-Performance Super-Resolution via Patch-Based Deep Neural Network for Real-Time Implementation" in IEICE TRANSACTIONS on Information,
vol. E101-D, no. 11, pp. 2808-2817, November 2018, doi: 10.1587/transinf.2018EDP7081.
Abstract: Recently, Super-resolution convolutional neural network (SRCNN) is widely known as a state of the art method for achieving single-image super resolution. However, performance problems such as jaggy and ringing artifacts exist in SRCNN. Moreover, in order to realize a real-time upconverting system for high-resolution video streams such as 4K/8K 60 fps, problems such as processing delay and implementation cost remain. In the present paper, we propose high-performance super-resolution via patch-based deep neural network (SR-PDNN) rather than a convolutional neural network (CNN). Despite the very simple end-to-end learning system, the SR-PDNN achieves higher performance than the conventional CNN-based approach. In addition, this system is suitable for ultra-low-delay video processing by hardware implementation using an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7081/_p
Copy
@ARTICLE{e101-d_11_2808,
author={Reo AOKI, Kousuke IMAMURA, Akihiro HIRANO, Yoshio MATSUDA, },
journal={IEICE TRANSACTIONS on Information},
title={High-Performance Super-Resolution via Patch-Based Deep Neural Network for Real-Time Implementation},
year={2018},
volume={E101-D},
number={11},
pages={2808-2817},
abstract={Recently, Super-resolution convolutional neural network (SRCNN) is widely known as a state of the art method for achieving single-image super resolution. However, performance problems such as jaggy and ringing artifacts exist in SRCNN. Moreover, in order to realize a real-time upconverting system for high-resolution video streams such as 4K/8K 60 fps, problems such as processing delay and implementation cost remain. In the present paper, we propose high-performance super-resolution via patch-based deep neural network (SR-PDNN) rather than a convolutional neural network (CNN). Despite the very simple end-to-end learning system, the SR-PDNN achieves higher performance than the conventional CNN-based approach. In addition, this system is suitable for ultra-low-delay video processing by hardware implementation using an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).},
keywords={},
doi={10.1587/transinf.2018EDP7081},
ISSN={1745-1361},
month={November},}
Copy
TY - JOUR
TI - High-Performance Super-Resolution via Patch-Based Deep Neural Network for Real-Time Implementation
T2 - IEICE TRANSACTIONS on Information
SP - 2808
EP - 2817
AU - Reo AOKI
AU - Kousuke IMAMURA
AU - Akihiro HIRANO
AU - Yoshio MATSUDA
PY - 2018
DO - 10.1587/transinf.2018EDP7081
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E101-D
IS - 11
JA - IEICE TRANSACTIONS on Information
Y1 - November 2018
AB - Recently, Super-resolution convolutional neural network (SRCNN) is widely known as a state of the art method for achieving single-image super resolution. However, performance problems such as jaggy and ringing artifacts exist in SRCNN. Moreover, in order to realize a real-time upconverting system for high-resolution video streams such as 4K/8K 60 fps, problems such as processing delay and implementation cost remain. In the present paper, we propose high-performance super-resolution via patch-based deep neural network (SR-PDNN) rather than a convolutional neural network (CNN). Despite the very simple end-to-end learning system, the SR-PDNN achieves higher performance than the conventional CNN-based approach. In addition, this system is suitable for ultra-low-delay video processing by hardware implementation using an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA).
ER -