2022
DOI: 10.1109/tgrs.2021.3113912
|View full text |Cite
|
Sign up to set email alerts
|

Asymmetric Siamese Networks for Semantic Change Detection in Aerial Images

Abstract: Given two multi-temporal aerial images, semantic change detection aims to locate the land-cover variations and identify their categories with pixel-wise boundaries. The problem has demonstrated promising potentials in many earth vision related tasks, such as precise urban planning and natural resource management. Existing state-of-the-art algorithms mainly identify the changed pixels through symmetric modules, which would suffer from categorical ambiguity caused by changes related to totally different landcove… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
46
0

Year Published

2022
2022
2023
2023

Publication Types

Select...
9
1

Relationship

0
10

Authors

Journals

citations
Cited by 86 publications
(46 citation statements)
references
References 61 publications
0
46
0
Order By: Relevance
“…Alternatively, SCD is treated as a semantic segmentation problem of changed and unchanged regions [11]. Yang et al [28] present an asymmetric Siamese network with heterogeneous feature extraction to alleviate the categorical ambiguity in the semantic identification process. Most of the works treat the localization of changed regions and identify the semantic classes separately.…”
Section: B Scd Algorithmsmentioning
confidence: 99%
“…Alternatively, SCD is treated as a semantic segmentation problem of changed and unchanged regions [11]. Yang et al [28] present an asymmetric Siamese network with heterogeneous feature extraction to alleviate the categorical ambiguity in the semantic identification process. Most of the works treat the localization of changed regions and identify the semantic classes separately.…”
Section: B Scd Algorithmsmentioning
confidence: 99%
“…N. images Tile size CD map Notes SZTAKI Air change [31,32] 13 952x640 2D DSIFN-CD [33] 3600 512x512 2D SECOND [34] 4662 512x512 2D 6 classes OSCD [35] 24 600x600 2D multispectral S2Looking [36] 5000 1024x1024 2D S2MTCP [37] 1520 600x600 2D SYSU-CD [38] 20000 256x256 2D 6 classes DynamicEarthNet [39] 54750 1024x1024 2D dense annotations LEVIR-CD [40] 637 1024x1024 2D 3DCD [5] (ours) 472 400x400 2D -3D…”
Section: Namementioning
confidence: 99%
“…Even though existing methods have made great achievements in remote sensing CD, there are still challenges to achieve finegrained cropland change detection. The performance of a deep learning model largely depends on the training dataset, and many previous works have provided well-annotated datasets for change detection, such as HRSCD [42] and SECOND [43] for semantic CD, SYSU-CD [44] and SVCD [45] for binary CD, and BCDD [46] and LEVIR-CD [31] for building CD, etc. So far, there is no dataset that specifically focuses on cropland changes, which greatly limits the development and application of cropland change detection models.…”
Section: Introductionmentioning
confidence: 99%