已入深夜,您辛苦了!由于当前在线用户较少,发布求助请尽量完整的填写文献信息,科研通机器人24小时在线,伴您度过漫漫科研夜!祝你早点完成任务,早点休息,好梦!

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
yyjdtc发布了新的文献求助30
1秒前
麦尔哈巴完成签到 ,获得积分10
3秒前
傲娇风格完成签到,获得积分10
3秒前
端庄向雁发布了新的文献求助10
3秒前
9秒前
Tim完成签到 ,获得积分10
12秒前
充电宝应助结实初翠采纳,获得10
13秒前
lgf完成签到,获得积分10
14秒前
14秒前
或习发布了新的文献求助10
15秒前
旅行者完成签到 ,获得积分10
15秒前
17秒前
Jasper应助或习采纳,获得10
21秒前
Soda完成签到,获得积分10
21秒前
干净菀完成签到,获得积分10
26秒前
闪闪善若完成签到 ,获得积分10
28秒前
30秒前
或习完成签到,获得积分10
32秒前
32秒前
潮人完成签到 ,获得积分10
33秒前
yyjdtc完成签到,获得积分10
33秒前
李李发布了新的文献求助10
35秒前
39秒前
草莓啵啵兔完成签到 ,获得积分10
40秒前
Jasper应助科研通管家采纳,获得10
55秒前
sumkki30应助科研通管家采纳,获得10
55秒前
上官若男应助科研通管家采纳,获得10
55秒前
科研通AI2S应助科研通管家采纳,获得10
55秒前
星辰大海应助科研通管家采纳,获得10
55秒前
NexusExplorer应助qifunongsuo1213采纳,获得10
56秒前
57秒前
1分钟前
缓慢的翅膀完成签到,获得积分10
1分钟前
summer夏发布了新的文献求助10
1分钟前
steven完成签到 ,获得积分10
1分钟前
田様应助钮水香采纳,获得10
1分钟前
善学以致用应助芮卜肆采纳,获得10
1分钟前
123完成签到,获得积分20
1分钟前
小马甲应助林博2025采纳,获得10
1分钟前
顺风顺水顺财神完成签到 ,获得积分10
1分钟前
高分求助中
Spray / Wall-interaction Modelling by Dimensionless Data Analysis 2000
Mixed-anion Compounds 600
A new species of Velataspis (Hemiptera Coccoidea Diaspididae) from tea in Assam 500
Earth System Geophysics 500
Aspects of Babylonian celestial divination: the lunar eclipse tablets of Enūma Anu Enlil 500
Geochemistry, 2nd Edition 地球化学经典教科书第二版 401
2024 Medicinal Chemistry Reviews 400
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 3200652
求助须知:如何正确求助?哪些是违规求助? 2850426
关于积分的说明 8071997
捐赠科研通 2514157
什么是DOI,文献DOI怎么找? 1346910
科研通“疑难数据库(出版商)”最低求助积分说明 640281
邀请新用户注册赠送积分活动 610407