Leveraging local and global descriptors in parallel to search correspondences for visual localization

Vision-based localization is the problem of inferring the pose of the camera given a single image. One commonly used approach relies on image retrieval where the query input is compared against a database of localized support examples and its pose is inferred with the help of the retrieved items. This assumes that images taken from the same places consist of the same landmarks and thus would have similar feature representations. These representations can learn to be robust to different variations in capture conditions like time of the day or weather. In this work, we introduce a framework which aims at enhancing the performance of such retrieval-based localization methods. It consists in taking into account additional information available, such as GPS coordinates or temporal proximity in the acquisition of the images. More precisely, our method consists in constructing a graph based on this additional information that is later used to improve reliability of the retrieval process by filtering the feature representations of support and/or query images. We show that the proposed method is able to significantly improve the localization accuracy on two large scale datasets, as well as the mean average precision in classical image retrieval scenarios.

Download Full-text

Visual Localization Based on Remote Sensing Scene Matching with Siamese Feature Aggregation Network

IGARSS 2020 - 2020 IEEE International Geoscience and Remote Sensing Symposium ◽

10.1109/igarss39084.2020.9323256 ◽

2020 ◽

Author(s):

Wang Chen ◽

Yuan Yuan ◽

Ganchao Liu

Keyword(s):

Remote Sensing ◽

Visual Localization ◽

Scene Matching ◽

Feature Aggregation

Download Full-text

Domain-invariant similarity activation map metric learning for retrieval-based long-term visual localization

IEEE/CAA Journal of Automatica Sinica ◽

10.1109/jas.2021.1003907 ◽

2021 ◽

pp. 1-16

Author(s):

Hanjiang Hu ◽

Hesheng Wang ◽

Zhe Liu ◽

Weidong Chen

Keyword(s):

Metric Learning ◽

Visual Localization ◽

Activation Map

Download Full-text

Leaf Plasmodesmata Respond Differently to TMV, ToBRFV and TYLCV Infection

Plants ◽

10.3390/plants10071442 ◽

2021 ◽

Vol 10 (7) ◽

pp. 1442

Author(s):

Yaarit Kutsher ◽

Dalia Evenor ◽

Eduard Belausov ◽

Moshe Lapidot ◽

Moshe Reuveni

Keyword(s):

Cell Types ◽

Plant Viruses ◽

Tomato Leaf ◽

Visual Localization ◽

Tobacco Leaf ◽

Signal Distribution ◽

Movement Proteins ◽

Significant Difference ◽

Intracellular Mechanism ◽

Leaf Curl Virus

Macromolecule and cytosolic signal distribution throughout the plant employs a unique cellular and intracellular mechanism called plasmodesmata (PD). Plant viruses spread throughout plants via PD using their movement proteins (MPs). Viral MPs induce changes in plasmodesmata’s structure and alter their ability to move macromolecule and cytosolic signals. The developmental distribution of a family member of proteins termed plasmodesmata located proteins number 5 (PDLP5) conjugated to GFP (PDLP5-GFP) is described here. The GFP enables the visual localization of PDLP5 in the cell via confocal microscopy. We observed that PDLP5-GFP protein is present in seed protein bodies and immediately after seed imbibition in the plasma membrane. The effect of three different plant viruses, the tobacco mosaic virus (TMV), tomato brown rugose fruit virus (ToBRFV, tobamoviruses), and tomato yellow leaf curl virus (TYLCV, begomoviruses), on PDLP5-GFP accumulation at the plasmodesmata was tested. In tobacco leaf, TMV and ToBRFV increased PDLP5-GFP amount at the plasmodesmata of cell types compared to control. However, there was no statistically significant difference in tomato leaf. On the other hand, TYLCV decreased PDLP5-GFP quantity in plasmodesmata in all tomato leaf cells compared to control, without any significant effect on plasmodesmata in tobacco leaf cells.

Download Full-text