ASLFeat: Learning Local Features of Accurate Shape and Localization

Luo, Zixin; Zhou, Lei; Bai, Xuyang; Chen, Hongkai; Zhang, Jiahui; Yao, Yao; Li, Shiwei; Fang, Tian; Quan, Long

doi:10.1109/cvpr42600.2020.00662

Cited by 286 publications

(227 citation statements)

References 30 publications

Supporting

Mentioning

226

Contrasting

Order By: Relevance

“…These are very interesting and can potentially lead to a better performance in terms of fewer outliers. On the other hand, as reported in , Luo et al (2020), Jin et al (2020) and Bojanić et al (2020), we should expect a very low keypoint localization accuracy or the extraction of a limited number of features (DeTone et al, 2018), which, in both cases, lead to a less accurate 3D reconstruction. Unlike similar investigations, we propose to evaluate learning-based methods with different metrics, using various image blocks (Table 1) and considering bundle adjustment statistics as well as 3D reference points (targets measured with topographic methods) as ground truth.…”

Section: Introductionmentioning

confidence: 69%

“…End-to-end. These approaches perform a simultaneous detection and description in order to extract recognizable and uniquely describable keypoints (Yi et al, 2016;DeTone et al, 2018;Ono et al, 2019;Dusmanu et al, 2019;Revaud et al, 2019;Christiansen et al, 2019;Luo et al, 2020). The joint training of descriptor and detector avoids extracting non-discriminative keypoints and selects only repeatable interest points to improve the overall feature matching pipeline.…”

Section: Learning-based Methodsmentioning

confidence: 99%

“…So researchers aimed to let a deep network discover automatically which feature extraction process and representation are most suited to the data (Revaud et al, 2019). Some end-to-end methods for image-based 3D reconstruction purposes under challenging conditions are now available and solutions based on traditional hand-crafted methods are beginning to be outperformed by state-of-the-art learning-based approaches (Yi et al, 2016;DeTone et al, 2018;Ono et al, 2019;Dusmanu et al, 2019;Revaud et al, 2019;Christiansen et al, 2019;Luo et al, 2020). Among the current limitations for these learning-based methods, we can list the lack of shape-awareness and the general invariance to geometric, radiometric and scale changes, low localization accuracy, repeatability, etc.…”

Section: Introductionmentioning

confidence: 99%

“…Among the current limitations for these learning-based methods, we can list the lack of shape-awareness and the general invariance to geometric, radiometric and scale changes, low localization accuracy, repeatability, etc. These issues can degrade the extractor performance in very different ways depending on how it has been designed (Revaud et al, 2019;Luo et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluating Hand-Crafted and Learning-Based Features for Photogrammetric Applications

Remondino

Menna

Morelli

2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. The image orientation (or Structure from Motion – SfM) process needs well localized, repeatable and stable tie points in order to derive camera poses and a sparse 3D representation of the surveyed scene. The accurate identification of tie points in large image datasets is still an open research topic in the photogrammetric and computer vision communities. Tie points are established by firstly extracting keypoint using a hand-crafted feature detector and descriptor methods. In the last years new solutions, based on convolutional neural network (CNN) methods, were proposed to let a deep network discover which feature extraction process and representation are most suitable for the processed images. In this paper we aim to compare state-of-the-art hand-crafted and learning-based method for the establishment of tie points in various and different image datasets. The investigation highlights the actual challenges for feature matching and evaluates selected methods under different acquisition conditions (network configurations, image overlap, UAV vs terrestrial, strip vs convergent) and scene's characteristics. Remarks and lessons learned constrained to the used datasets and methods are provided.

show abstract

Section: Introductionmentioning

confidence: 69%

Section: Learning-based Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evaluating Hand-Crafted and Learning-Based Features for Photogrammetric Applications

Remondino

Menna

Morelli

2021

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Key.Net [ 25 ] combines the hand-crafted detector and CNN filter in the shallow multi-scale framework, which reduces the number of network parameters and ensures the detection repeatibility rate. ASLFeat [ 26 ] further improves the positioning accuracy of D2-Net keypoints.…”

Section: Introductionmentioning

confidence: 99%

Illumination-Invariant Feature Point Detection Based on Neighborhood Information

Wang

Zeng

et al. 2020

Sensors

View full text Add to dashboard Cite

Feature point detection is the basis of computer vision, and the detection methods with geometric invariance and illumination invariance are the key and difficult problem in the field of feature detection. This paper proposes an illumination-invariant feature point detection method based on neighborhood information. The method can be summarized into two steps. Firstly, the feature points are divided into eight types according to the number of connected neighbors. Secondly, each type of feature points is classified again according to the position distribution of neighboring pixels. The theoretical deduction proves that the proposed method has lower computational complexity than other methods. The experimental results indicate that, when the photometric variation of the two images is very large, the feature-based detection methods are usually inferior, while the learning-based detection methods performs better. However, our method performs better than the learning-based detection method in terms of the number of feature points, the number of matching points, and the repeatability rate stability. The experimental results demonstrate that the proposed method has the best illumination robustness among state-of-the-art feature detection methods.

show abstract