Power Efficient Video Super-Resolution on Mobile NPUs with Deep Learning, Mobile AI &amp; AIM 2022 Challenge: Report

Ignatov, Andrey; Timofte, Radu; Chiang, Cheng-Ming; Kuo, Hsien-Kai; Xu, Yulong; Man-Yu, Lee,; Lu, Amanda; Cheng, Chia-Ming; Chen, Chih‐Cheng; Jia-Ying, Yong,; Shuai, Hong-Han; Cheng, Wen-Huang; Jia, Zhuang; Xu, Tianyu; Zhang, Yijian; Bao, Long; Sun, Hao; Zhang, Diankai; Gao, Si; Liu, Shaoli; Wu, Biao; Zhang, Xiaofeng; Zheng, Chengjian; Lu, Kaidi; Wang, Ning; Sun, Xiao; Wu, Haodong; Liu, Xuncheng; Zhang, Weizhan; Cheng, Yan; Du, Haipeng; Zheng, Qinghua; Wang, Qi; Chen, Wangdu; Duan, Ran; Sun, Mengdi; Zhu, Dan; Chen, Guannan; Cho, Hojin; Kim, Steve; Yue, Shijie; Li, Chenghua; Zhuge, Zhengyang; Chen, Wei; Wang, Wen-Xu; Zhou, Yue; Cai, Xiaojun; Cai, Hengxing; Xu, Kele; Liu, Li; Cheng, Zehua; Lian, Wenyi; Lian, Wenjing

doi:10.1007/978-3-031-25066-8_6

Cited by 2 publications

(2 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A avaliac ¸ão dos modelos será baseada na metodologia proposta em [14] para a Mobile AI & AIM 2022 Challenge, a competic ¸ão focada em avaliar modelos de redes neurais profundas em smartphones. Os principais critérios são fidelidade da imagem resultante, percepc ¸ão e tempo de execuc ¸ão, e tais critérios são avaliados através da ferramenta AI Benchmark 1 , um aplicativo móvel para smartphones que utilizam o sistema operacional Android.…”

Section: B Medidas De Avaliac ¸ãOunclassified

“…O terceiro critério é o tempo de execuc ¸ão, medido em milissegundos, usando tanto a CPU quanto a GPU do smartphone. Com isso o resultado final é obtido pela Equac ¸ão 1, baseada na equac ¸ão desenvolvida por [14]. O consumo de energia, que não foi possível de ser medido, foi substituído pelo tempo de execuc ¸ão T E dividido por δ (o tempo dos quadros em uma taxa de quadros de 30 FPS, ou seja 33, 3ms).…”

Section: B Medidas De Avaliac ¸ãOunclassified

See 1 more Smart Citation

Generative Adversarial Network and ResNet Comparison for Video Super Resolution in Smartphones

Vitória,

Boldt,

Seibel Junior

2023

Anais Do XVIII Workshop De Visão Computacional (WVC 2023)

View full text Add to dashboard Cite

With the latest evolutions of smartphone hardware regarding neural network processing and advances in Image Super Resolution, one question remains: What is the applicability of SR techniques already consolidated in smartphones? Several works focused on different characteristics of images have been developed and seek to understand how such networks can meet different segments. This work focuses on applying two already consolidated networks in the processing of photo-realistic images, SRResNet and SRGAN, in mobile devices to identify how such networks behave in smartphones of different generations, either in the image quality or in the response time of these networks. After evaluation, the SRResNet network had a better performance both in inferred image quality and lower latency, with a PSNR of 27.7075 versus 21.3843 and 0.19 milliseconds latency, compared to 0.20 milliseconds for SRGAN, thus showing that it is feasible to apply SR techniques already consolidated in smartphones.

show abstract

Section: B Medidas De Avaliac ¸ãOunclassified