Using multiple optical cameras for correspondence identification between objects in the fields of view

Roman N. Kvyetnyy, Volodymyr Kotsiubynskyi, Oleksandr Kyrylenko, Oleg K. Kolesnytskyj, Viktoria Dumenko, Andrzej Kotyra, Dinara Mussayeva, Nazerke Abilkaiyr

2023

Full text Open access: Green Metadata (CSL-JSON)
DOI 10.1117/12.3022791
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

The study aims to explore a method for identifying corresponding objects across multiple camera views, to improve the accuracy of object re-identification. We analyzed various techniques, including contour detection, region of interest extraction, and keypoint extraction. We also examined the challenges of finding object correspondences between multiple camera views. To evaluate the effectiveness of the proposed method, we utilized two human attribute datasets, Market-1501 and DukeMTMC-reID, and performed extensive testing on these datasets.