ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2019
DOI: 10.1109/icassp.2019.8683541
|View full text |Cite
|
Sign up to set email alerts
|

DSSLIC: Deep Semantic Segmentation-based Layered Image Compression

Abstract: Deep learning has revolutionized many computer vision fields in the last few years, including learning-based image compression. In this paper, we propose a deep semantic segmentation-based layered image compression (DSSLIC) framework in which the semantic segmentation map of the input image is obtained and encoded as the base layer of the bit-stream. A compact representation of the input image is also generated and encoded as the first enhancement layer. The segmentation map and the compact version of the imag… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

1
73
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
5
3
2

Relationship

1
9

Authors

Journals

citations
Cited by 90 publications
(74 citation statements)
references
References 26 publications
1
73
0
Order By: Relevance
“…However, in the indoor environment, mutual positioning based on unknown environment is still a relatively new field, and there are not many reference materials. Therefore, this paper starts from the research status of visual positioning technology and image semantic segmentation technology [12][13][14] and proposes a method for mutual positioning between users in an unknown environment. In this paper, a visual positioning method based on identification is adopted to solve the user's position coordinates in the current coordinate system by taking the same object that can be seen between users as the center of the coordinate system.…”
Section: Related Workmentioning
confidence: 99%
“…However, in the indoor environment, mutual positioning based on unknown environment is still a relatively new field, and there are not many reference materials. Therefore, this paper starts from the research status of visual positioning technology and image semantic segmentation technology [12][13][14] and proposes a method for mutual positioning between users in an unknown environment. In this paper, a visual positioning method based on identification is adopted to solve the user's position coordinates in the current coordinate system by taking the same object that can be seen between users as the center of the coordinate system.…”
Section: Related Workmentioning
confidence: 99%
“…Despite that the structure representation can be well analyzed, the support for vision tasks, such as image synthesis and modification, are limited by pixel-level color representations. The semantic guidance is used to assist in reconstructing images from compact compressed visual data in [14], [27], where the main data stream is still signal-oriented though.…”
Section: Conceptual Compressionmentioning
confidence: 99%
“…However, the improvement of visual quality is not exactly equivalent to reducing signal distortion in the conventional sense. The GAN-based model [18,19] output is visually realistic, but their objective indicators such as PSNR/SSIM are not satisfying. Besides, the GAN output's visual quality is measured by the VGG network, FID, GAN discriminator, user survey, etc.…”
Section: Introduction and Previous Workmentioning
confidence: 99%