Unsupervised Global and Local Homography Estimation With Motion Basis Learning

Shuaicheng, Liu,; Lu, Yuhang; Hai, Jiang; Ye, Nianjin; Wang, Chuan; Zeng, Bing

doi:10.1109/tpami.2022.3223789

Cited by 8 publications

(5 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This may be because that our method estimates H from local regions, which is affected by low texture regions in these scenes. When compared with multi-H methods (Row 12-15), the SOTA method is MeshBasesHomo (Liu et al 2022a). Our method outperforms it by 13% (0.79→0.69) when using similar amount of H (our method uses at most 4 H).…”

Section: Comparison With Existing Methodsmentioning

confidence: 88%

“…Dataset Our method is evaluated on a natural image dataset (Zhang et al 2020;Liu et al 2022a) with 75.8k training pairs and 4.2k testing pairs. The scenes in the dataset are roughly categorized into five types: REgular (RE), Low Texture (LT), Low Light (LL), Small Foreground (SF) and Large Foreground (LF), where the last four types are more challenging.…”

Section: Methodsmentioning

confidence: 99%

“…Recently, (Liu et al 2016) and (Nie et al 2022) propose to estimate multiple Hs by dividing the image into mesh grids and computing a local H for each grid. (Liu et al 2022b) and (Liu et al 2022a) also propose to use mesh grid H estimation. By dividing the image into even mesh grids, there will be both cases where one grid contains multiple…”

Section: Introductionmentioning

confidence: 99%

“…Another issue is that when evaluating H estimation performance, most existing works use Mean Squared Error (MSE), also referred to as Point Matching Error in (Zhang et al 2020;Hong et al 2022;Liu et al 2022a). MSE calculates the deviations between the labeled matching point pairs within image pairs.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Mask-Homo: Pseudo Plane Mask-Guided Unsupervised Multi-Homography Estimation

Wang,

Liu,

Zhang

et al. 2024

AAAI

View full text Add to dashboard Cite

Homography estimation is a fundamental problem in computer vision. Previous works mainly focus on estimating either a single homography, or multiple homographies based on mesh grid division of the image. In practical scenarios, single homography is inadequate and often leads to a compromised result for multiple planes; while mesh grid multi-homography damages the plane distribution of the scene, and does not fully address the restriction to use homography. In this work, we propose a novel semantics guided multi-homography estimation framework, Mask-Homo, to provide an explicit solution to the multi-plane depth disparity problem. First, a pseudo plane mask generation module is designed to obtain multiple correlated regions that follow the plane distribution of the scene. Then, multiple local homography transformations, each of which aligns a correlated region precisely, are predicted and corresponding warped images are fused to obtain the final result. Furthermore, a new metric, Mask-PSNR, is proposed for more comprehensive evaluation of alignment. Extensive experiments are conducted to verify the effectiveness of the proposed method. Our code is available at https://github.com/SAITPublic/MaskHomo.

show abstract

Section: Comparison With Existing Methodsmentioning

confidence: 88%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Mask-Homo: Pseudo Plane Mask-Guided Unsupervised Multi-Homography Estimation

Wang,

Liu,

Zhang

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…As illustrated by Fig. 2(b), we follow existing works (Babbar and Bajaj 2022;Ye et al 2021;Liu et al 2022b) to simulate the dynamic changes of k t during the motion by computing the inter-frame Homography matrix H t−1→t , formulated as:…”

Section: Problem Formulation and Datasetmentioning

confidence: 99%

Decoupling Degradations with Recurrent Network for Video Restoration in Under-Display Camera

Liu,

Wang,

Fan

et al. 2024

AAAI

View full text Add to dashboard Cite

Under-display camera (UDC) systems are the foundation of full-screen display devices in which the lens mounts under the display. The pixel array of light-emitting diodes used for display diffracts and attenuates incident light, causing various degradations as the light intensity changes. Unlike general video restoration which recovers video by treating different degradation factors equally, video restoration for UDC systems is more challenging that concerns removing diverse degradation over time while preserving temporal consistency. In this paper, we introduce a novel video restoration network, called D2RNet, specifically designed for UDC systems. It employs a set of Decoupling Attention Modules (DAM) that effectively separate the various video degradation factors. More specifically, a soft mask generation function is proposed to formulate each frame into flare and haze based on the diffraction arising from incident light of different intensities, followed by the proposed flare and haze removal components that leverage long- and short-term feature learning to handle the respective degradations. Such a design offers an targeted and effective solution to eliminating various types of degradation in UDC systems. We further extend our design into multi-scale to overcome the scale-changing of degradation that often occur in long-range videos. To demonstrate the superiority of D2RNet, we propose a large-scale UDC video benchmark by gathering HDR videos and generating realistically degraded videos using the point spread function measured by a commercial UDC system. Extensive quantitative and qualitative evaluations demonstrate the superiority of D2RNet compared to other state-of-the-art video restoration and UDC image restoration methods.

show abstract