2021
DOI: 10.1145/3419842
|View full text |Cite
|
Sign up to set email alerts
|

A Weakly Supervised Semantic Segmentation Network by Aggregating Seed Cues: The Multi-Object Proposal Generation Perspective

Abstract: Weakly supervised semantic segmentation under image-level annotations is effectiveness for real-world applications. The small and sparse discriminative regions obtained from an image classification network that are typically used as the important initial location of semantic segmentation also form the bottleneck. Although deep convolutional neural networks (DCNNs) have exhibited promising performances for single-label image classification tasks, images of the real-world usually contain multiple categories, whi… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
19
0

Year Published

2021
2021
2023
2023

Publication Types

Select...
10

Relationship

0
10

Authors

Journals

citations
Cited by 60 publications
(19 citation statements)
references
References 28 publications
0
19
0
Order By: Relevance
“…(1) The original type C is classified as type C, and the quantity is recorded as a [36,37] visualizes the layers of CNN and shows the above theory more clearly. The bottom layer 1 and 2 of the model network can usually learn the basic color and edge features of the object; the third layer of the model network generally learns the texture features of the object; the fourth layer that continues upward can learn local features, such as wheels; the top level learns more discernible overall characteristics.…”
Section: Experimental Settingsmentioning
confidence: 99%
“…(1) The original type C is classified as type C, and the quantity is recorded as a [36,37] visualizes the layers of CNN and shows the above theory more clearly. The bottom layer 1 and 2 of the model network can usually learn the basic color and edge features of the object; the third layer of the model network generally learns the texture features of the object; the fourth layer that continues upward can learn local features, such as wheels; the top level learns more discernible overall characteristics.…”
Section: Experimental Settingsmentioning
confidence: 99%
“…Specifically, Res2Net implements multiscale features via splitting channels of feature maps into subgroups and fusing these channel groups hierarchically. It has been proven that Res2Net can boost many backbone networks in some vision tasks, including object detection, semantic segmentation [32,33], and salient object detection [29].…”
Section: High Spatial Resolutionmentioning
confidence: 99%
“…The distant supervision hypothesis means that when there is a Wireless Communications and Mobile Computing relationship between two entities, then all sentences containing the pair of entities are considered to express this relationship to some extent. Distant supervision is to provide labels for data with the help of external knowledge bases, to save the trouble of manual labelling [37]. Attributes can also be considered as a type of relationship, so the distant supervision assumption is applied to the annotation of attribute data.…”
Section: Data Annotationmentioning
confidence: 99%