Self-Supervised Learning for RGB-Guided Depth Enhancement by Exploiting the Dependency Between RGB and Depth

RGB颜色模型 人工智能 计算机科学 依赖关系(UML) 计算机视觉 降噪 噪音(视频) 模式识别(心理学) 图像(数学)
作者
Jun Wang,Peilin Liu,Fei Wen
出处
期刊:IEEE transactions on image processing [Institute of Electrical and Electronics Engineers]
卷期号:32: 159-174 被引量:5
标识
DOI:10.1109/tip.2022.3226419
摘要

Due to the imaging mechanism of time-of-flight (ToF) sensors, the captured depth images usually suffer from severe noise and degradation. Though many RGB-guided methods have been proposed for depth image enhancement in the past few years, yet the enhancement performance on real-world depth images is still largely unsatisfactory. Two main reasons are the complexity of realistic noise and degradation in depth images, and the difficulty in collecting noise-clean pairs for supervised enhancement learning. This work aims to develop a self-supervised learning method for RGB-guided depth image enhancement, which does not require any noisy-clean pairs but can significantly boost the enhancement performance on real-world noisy depth images. To this end, we exploit the dependency between RGB and depth images to self-supervise the learning of the enhancement model. It is achieved by maximizing the cross-modal dependency between RGB and depth to promote the enhanced depth having dependency with the RGB of the same scene as much as possible. Furthermore, we augment the cross-modal dependency maximization formulation based on the optimal transport theory to achieve further performance improvement. Experimental results on both synthetic and real-world data demonstrate that our method can significantly outperform existing state-of-the-art methods on depth denoising, multi-path interference suppression, and hole filling. Particularly, our method shows remarkable superiority over existing ones on real-world data in handling various realistic complex degradation. Code is available at https://github.com/wjcyt/SRDE.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
大幅提高文件上传限制,最高150M (2024-4-1)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
ywzwszl完成签到,获得积分10
刚刚
Mottri完成签到 ,获得积分10
刚刚
疯狂吃辣完成签到 ,获得积分10
1秒前
浅香千雪发布了新的文献求助10
1秒前
所所应助青木蓝采纳,获得10
1秒前
Ava应助wst采纳,获得10
1秒前
宇婷发布了新的文献求助10
1秒前
独特的板凳完成签到,获得积分10
1秒前
默默小鸽子完成签到,获得积分10
2秒前
2秒前
Hyacinth完成签到,获得积分10
3秒前
悬搁宁静完成签到,获得积分10
3秒前
3秒前
3秒前
3秒前
111完成签到,获得积分20
3秒前
爆米花应助忧郁的犀牛采纳,获得10
3秒前
和谐幻柏完成签到,获得积分10
3秒前
yufanhui应助Harry采纳,获得10
4秒前
sh发布了新的文献求助10
4秒前
Eva完成签到,获得积分10
4秒前
上官若男应助小余采纳,获得10
4秒前
波喔博发布了新的文献求助10
4秒前
sally应助black的hole采纳,获得10
5秒前
Cold-Drink-Shop完成签到,获得积分10
5秒前
CipherSage应助ttt采纳,获得10
6秒前
搜集达人应助研友_LNB5DL采纳,获得10
7秒前
优秀的寄灵完成签到,获得积分10
7秒前
材料若饥完成签到,获得积分10
7秒前
小柚子it发布了新的文献求助10
8秒前
karen发布了新的文献求助10
8秒前
9秒前
JamesPei应助zjy03259采纳,获得10
9秒前
Yolanda完成签到 ,获得积分10
9秒前
Voyager完成签到,获得积分10
9秒前
10秒前
pipipi完成签到,获得积分10
10秒前
τ涛完成签到,获得积分10
11秒前
Shelley发布了新的文献求助30
11秒前
科目三应助Rain采纳,获得10
11秒前
高分求助中
Evolution 3rd edition 1500
保险藏宝图 1000
Lire en communiste 1000
Mantiden: Faszinierende Lauerjäger Faszinierende Lauerjäger 700
PraxisRatgeber: Mantiden: Faszinierende Lauerjäger 700
Mathematics and Finite Element Discretizations of Incompressible Navier—Stokes Flows 500
A new species of Coccus (Homoptera: Coccoidea) from Malawi 500
热门求助领域 (近24小时)
化学 医学 生物 材料科学 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 基因 遗传学 催化作用 物理化学 免疫学 量子力学 细胞生物学
热门帖子
关注 科研通微信公众号,转发送积分 3183813
求助须知:如何正确求助?哪些是违规求助? 2834105
关于积分的说明 7997564
捐赠科研通 2496385
什么是DOI,文献DOI怎么找? 1332143
科研通“疑难数据库(出版商)”最低求助积分说明 636519
邀请新用户注册赠送积分活动 603736