数学建模社区-数学中国
标题: Transferred Multi-Perception Attention Networks for Remote Sensing Image Supe... [打印本页]
作者: 杨利霞 时间: 2020-11-13 16:10
标题: Transferred Multi-Perception Attention Networks for Remote Sensing Image Supe...
Transferred Multi-Perception Attention Networks for
( {; g- P. k4 J1 V4 m0 k
Remote Sensing Image Super-Resolution
+ L8 f& z" L- C$ L4 d0 ^
9 I' e2 a1 ^; j' u6 E- X; `, l; B0 W+ \: G2 o
$ u# O/ w8 p* r! j' _0 `" a
Image super-resolution (SR) reconstruction plays a key role in coping with the increasing
' d I: ~- A2 j5 C8 c4 Q- @$ v4 M( x& o* fdemand on remote sensing imaging applications with high spatial resolution requirements. Though7 W/ w( J4 W3 p$ f+ h
many SR methods have been proposed over the last few years, further research is needed to improve/ ^$ t' ~3 D; r# o$ r4 I. Q, v
SR processes with regard to the complex spatial distribution of the remote sensing images and the
8 V7 U. a7 ?% r9 _+ Udiverse spatial scales of ground objects. In this paper, a novel multi-perception attention network
4 e1 O9 i5 J+ ?5 ^(MPSR) is developed with performance exceeding those of many existing state-of-the-art models.
, A- K( }, j1 G. n2 tBy incorporating the proposed enhanced residual block (ERB) and residual channel attention group6 s: T2 l( L6 U$ E% B. e2 i8 [# O
(RCAG), MPSR can super-resolve low-resolution remote sensing images via multi-perception learning: b" N; j& d# W' o" E/ v2 W9 Z' s
and multi-level information adaptive weighted fusion. Moreover, a pre-train and transfer learning& a' F& T2 S7 R1 c4 c
strategy is introduced, which improved the SR performance and stabilized the training procedure.8 G8 E& ]' b/ V& X
Experimental comparisons are conducted using 13 state-of-the-art methods over a remote sensing
[6 J/ _8 A R/ z2 ddataset and benchmark natural image sets. The proposed model proved its excellence in both objective
$ Y: ~- K% J2 @$ T* N0 E6 Hcriterion and subjective perspective.# m0 r' S Q$ I# d8 ~6 F- j9 ^7 U
) x: t: t* R! ~! \, S- k5 K9 M$ L3 [! n! x0 o* r. o
-
-
Transferred Multi-Perception Attention Networks for.pdf
11.42 MB, 下载次数: 0, 下载积分: 体力 -2 点
| 欢迎光临 数学建模社区-数学中国 (http://www.madio.net/) |
Powered by Discuz! X2.5 |