Efficient and robust large-scale structure-from-motion via track selection and camera prioritization

Cui, Hainan; Shen, Shuhan; Gao, Wei; Liu, Hong‐Min; Wang, Zhiheng

doi:10.1016/j.isprsjprs.2019.08.005

Cited by 25 publications

(7 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From a macro point of view, the composition of the CAD system is very similar to the process of early clinical diagnosis and screening of fundus diseases by ophthalmologists by observing fundus images. If the whole process of early screening methods is compared to an assembly line, then at the front end of the pipeline, CAD can automatically perceive the distortion of the fundus image [13] and perform corresponding image enhancement [14], [36] and restoration [15], [35]; In the intermediate stage, CAD can segment the lesion and extract many features; At the back end , CAD can transform complex diagnostic logic into classification or clustering problems in machine logic, and classification or clustering problems can be solved by machine learning. Like CAD, image preprocessing and then image segmentation are performed [16], [17], then feature extraction, and finally the process of intelligent diagnosis through machine learning is actually imitating the doctor's diagnostic thinking.…”

Section: Related Work a Traditional Methodsmentioning

confidence: 99%

Multi-Label Classification of Fundus Images With EfficientNet

et al. 2020

View full text Add to dashboard Cite

Convolutional neural network (CNN) has achieved remarkable success in the field of fundus images due to its powerful feature learning ability. Computer-aided diagnosis can obtain information with reference value for doctors in clinical diagnosis or screening through proper processing and analysis of fundus images. However, most of the previous studies have focused on the detection of a certain fundus disease, and the simultaneous diagnosis of multiple fundus diseases still faces great challenges. We propose a multi-label classification ensemble model of fundus images based on CNN to directly detect one or more fundus diseases in the retinal fundus images. Every single model consists of two parts. The first part is a feature extraction network based on EfficientNet, and the second part is a custom classification neural network for multi-label classification problems. Finally, the output probabilities of different models are fused as the final recognition result. And it was trained and tested on the data set provided by ODIR 2019 (Peking University International Competition on Ocular Disease Intelligent Recognition). The experimental results show that our model can be trained on fewer data sets and get good results.

show abstract

Section: Related Work a Traditional Methodsmentioning

confidence: 99%

Multi-Label Classification of Fundus Images With EfficientNet

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Similarly in (Zhu et al 2018), images are divided into multiple partitions first and a global motion averaging is solved to determines cameras at partition boundaries. In (Cui et al 2019), the orthogonal maximum spanning tree (Cui et al 2018) is used to select relative geometry inliers. Shonan-based RA (Dellaert et al 2020) is proposed to recover globally optimal solutions under mild assumptions on the measurement noise.…”

Section: Related Workmentioning

confidence: 99%

MMA: Multi-Camera Based Global Motion Averaging

Cui

Shen

2022

AAAI

Self Cite

View full text Add to dashboard Cite

In order to fully perceive the surrounding environment, many intelligent robots and self-driving cars are equipped with a multi-camera system. Based on this system, the structure-from-motion (SfM) technology is used to realize scene reconstruction, but the fixed relative poses between cameras in the multi-camera system are usually not considered. This paper presents a tailor-made multi-camera based motion averaging system, where the fixed relative poses are utilized to improve the accuracy and robustness of SfM. Our approach starts by dividing the images into reference images and non-reference images, and edges in view-graph are divided into four categories accordingly. Then, a multi-camera based rotating averaging problem is formulated and solved in two stages, where an iterative re-weighted least squares scheme is used to deal with outliers. Finally, a multi-camera based translation averaging problem is formulated and a l1-norm based optimization scheme is proposed to compute the relative translations of multi-camera system and reference camera positions simultaneously. Experiments demonstrate that our algorithm achieves superior accuracy and robustness on various data sets compared to the state-of-the-art methods.

show abstract

“…Here, | • | counts the number of vertices. The second MST is extracted from G , and this ensures there are no repeated edges between two MSTs, i.e., so-called orthogonal MSTs [25,35]. The above processes are repeated several times, and these extracted orthogonal MSTs compose a seed view-graph G seed with N seed iterations.…”

Section: Initializationmentioning

confidence: 99%

“…Recent developments of SfM focus on large-scale image sets, such as internet photo collections [21][22][23] and UAV image sets [18,24]. They typically consider moderate match pairs and pay less attention to those due to repetitive structures [8,25,26] and very short baselines [27][28][29]. These problematic match pairs can degrade SfM reconstruction results, or even lead to failure.…”

Section: Introductionmentioning

confidence: 99%

Progressive Structure from Motion by Iteratively Prioritizing and Refining Match Pairs

Xiao

Qingsong

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Structure from motion (SfM) has been treated as a mature technique to carry out the task of image orientation and 3D reconstruction. However, it is an ongoing challenge to obtain correct reconstruction results from image sets consisting of problematic match pairs. This paper investigated two types of problematic match pairs, stemming from repetitive structures and very short baselines. We built a weighted view-graph based on all potential match pairs and propose a progressive SfM method (PRMP-PSfM) that iteratively prioritizes and refines its match pairs (or edges). The method has two main steps: initialization and expansion. Initialization is developed for reliable seed reconstruction. Specifically, we prioritize a subset of match pairs by the union of multiple independent minimum spanning trees and refine them by the idea of cycle consistency inference (CCI), which aims to infer incorrect edges by analyzing the geometric consistency over cycles of the view-graph. The seed reconstruction is progressively expanded by iteratively adding new minimum spanning trees and refining the corresponding match pairs, and the expansion terminates when a certain completeness of the block is achieved. Results from evaluations on several public datasets demonstrate that PRMP-PSfM can successfully accomplish the image orientation task for datasets with repetitive structures and very short baselines and can obtain better or similar accuracy of reconstruction results compared to several state-of-the-art incremental and hierarchical SfM methods.

show abstract

Efficient and robust large-scale structure-from-motion via track selection and camera prioritization

Cited by 25 publications

References 14 publications

Multi-Label Classification of Fundus Images With EfficientNet

Multi-Label Classification of Fundus Images With EfficientNet

MMA: Multi-Camera Based Global Motion Averaging

Progressive Structure from Motion by Iteratively Prioritizing and Refining Match Pairs

Contact Info

Product

Resources

About