WO2022214821A3 - Monocular depth estimation - Google Patents

Monocular depth estimation Download PDF

Info

Publication number
WO2022214821A3
WO2022214821A3 PCT/GB2022/050881 GB2022050881W WO2022214821A3 WO 2022214821 A3 WO2022214821 A3 WO 2022214821A3 GB 2022050881 W GB2022050881 W GB 2022050881W WO 2022214821 A3 WO2022214821 A3 WO 2022214821A3
Authority
WO
WIPO (PCT)
Prior art keywords
candidate
training image
training
depth
depth map
Prior art date
Application number
PCT/GB2022/050881
Other languages
French (fr)
Other versions
WO2022214821A2 (en
Inventor
Jostyn Biebele FUBARA
Liangchuan GU
Original Assignee
Robok Limited
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Robok Limited filed Critical Robok Limited
Publication of WO2022214821A2 publication Critical patent/WO2022214821A2/en
Publication of WO2022214821A3 publication Critical patent/WO2022214821A3/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle

Abstract

A computer-implemented method of training a depth estimation model to generate a depth map from a monocular image includes receiving first and second training images respectively of first and second views of a scene, processing the first training image using the depth estimation model to generate a candidate depth map comprising an array of candidate depth values, projecting the second training image using the candidate depth map to generate a candidate reconstruction of the first training image, and updating the depth estimation model so as to reduce a value of a loss function comprising a photometric difference term penalising a photometric difference between the first training image and the candidate reconstruction of the first training image dependent. The loss function further comprises a reflection mitigation term penalising a second derivative in the horizontal direction of at least a portion of the candidate depth map.
PCT/GB2022/050881 2021-04-07 2022-04-07 Monocular depth estimation WO2022214821A2 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB2104949.9 2021-04-07
GB2104949.9A GB2605621A (en) 2021-04-07 2021-04-07 Monocular depth estimation

Publications (2)

Publication Number Publication Date
WO2022214821A2 WO2022214821A2 (en) 2022-10-13
WO2022214821A3 true WO2022214821A3 (en) 2022-11-17

Family

ID=75883606

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/GB2022/050881 WO2022214821A2 (en) 2021-04-07 2022-04-07 Monocular depth estimation

Country Status (2)

Country Link
GB (1) GB2605621A (en)
WO (1) WO2022214821A2 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111274939B (en) * 2020-01-19 2023-07-14 交信北斗科技有限公司 Automatic extraction method for road pavement pothole damage based on monocular camera

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200258249A1 (en) * 2017-11-15 2020-08-13 Google Llc Unsupervised learning of image depth and ego-motion prediction neural networks

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103337066B (en) 2013-05-27 2016-05-18 清华大学 3D obtains the calibration steps of system
EP3040941B1 (en) 2014-12-29 2017-08-02 Dassault Systèmes Method for calibrating a depth camera
US10154176B1 (en) 2017-05-30 2018-12-11 Intel Corporation Calibrating depth cameras using natural objects with expected shapes
CN109559349B (en) 2017-09-27 2021-11-09 虹软科技股份有限公司 Method and device for calibration
US10628968B1 (en) 2018-12-05 2020-04-21 Toyota Research Institute, Inc. Systems and methods of calibrating a depth-IR image offset
CN209541744U (en) 2019-04-26 2019-10-25 昆明理工大学 A kind of caliberating device that photogrammetric post-processing auxiliary scale restores and orients

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200258249A1 (en) * 2017-11-15 2020-08-13 Google Llc Unsupervised learning of image depth and ego-motion prediction neural networks

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Computer Vision - ECCV 2020 Workshops : Glasgow, UK, August 23-28, 2020, Proceedings, Part V", vol. 12365, 21 July 2020 (2020-07-21), Cham, pages 582 - 600, XP055872919, ISSN: 0302-9743, ISBN: 978-3-030-68238-5, Retrieved from the Internet <URL:https://arxiv.org/pdf/2007.06936.pdf> DOI: 10.1007/978-3-030-58565-5_35 *
MAARTEN SCHELLEVIS: "Improving Self-Supervised Single View Depth Estimation by Masking Occlusion", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 29 August 2019 (2019-08-29), XP081489269 *
REZA MAHJOURIAN ET AL: "Unsupervised Learning of Depth and Ego-Motion from Monocular Video Using 3D Geometric Constraints", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 15 February 2018 (2018-02-15), XP080856755 *

Also Published As

Publication number Publication date
GB202104949D0 (en) 2021-05-19
WO2022214821A2 (en) 2022-10-13
GB2605621A (en) 2022-10-12

Similar Documents

Publication Publication Date Title
CN107862293B (en) Radar color semantic image generation system and method based on countermeasure generation network
US20200265597A1 (en) Method for estimating high-quality depth maps based on depth prediction and enhancement subnetworks
US20210042929A1 (en) Three-dimensional object detection method and system based on weighted channel features of a point cloud
WO2020092477A8 (en) Display engine for post-rendering processing
CN109308719B (en) Binocular parallax estimation method based on three-dimensional convolution
CN110570457B (en) Three-dimensional object detection and tracking method based on stream data
CN110689562A (en) Trajectory loop detection optimization method based on generation of countermeasure network
CN104299220B (en) A kind of method that cavity in Kinect depth image carries out real-time filling
US9807372B2 (en) Focused image generation single depth information from multiple images from multiple sensors
CN103796001A (en) Method and device for synchronously acquiring depth information and color information
CN103914813A (en) Colorful haze image defogging and illumination compensation restoration method
CN104077808A (en) Real-time three-dimensional face modeling method used for computer graph and image processing and based on depth information
AU2018200055A1 (en) Show personalized preview for 360 or panoramic images based on search query to enhance customer experience by providing them most relevant view of images
CN112561996A (en) Target detection method in autonomous underwater robot recovery docking
CN111354077A (en) Three-dimensional face reconstruction method based on binocular vision
CN114067197A (en) Pipeline defect identification and positioning method based on target detection and binocular vision
CN106952304A (en) A kind of depth image computational methods of utilization video sequence interframe correlation
WO2022214821A3 (en) Monocular depth estimation
Łuczyński et al. Underwater image haze removal with an underwater-ready dark channel prior
CN115375581A (en) Dynamic visual event stream noise reduction effect evaluation method based on event time-space synchronization
US20160225159A1 (en) Image processing method and image processing apparatus
CN103716615A (en) 2D video three-dimensional method based on sample learning and depth image transmission
CN109218706B (en) Method for generating stereoscopic vision image from single image
CN114549297A (en) Unsupervised monocular depth estimation method based on uncertain analysis
KR20110112143A (en) A method for transforming 2d video to 3d video by using ldi method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22722331

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE