Content-Based Multi-Camera Video Alignment using Accelerometer Data

Nazare, Antonio C.; Costa, Filipe; Schwartz, William Robson

doi:10.1109/avss.2018.8639468

Cited by 1 publication

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are two main approaches for maintaining tracks between cameras—a camera topology-based approach that identifies the sources and sinks of each camera view and tracks moving objects within camera views primarily based on motion and kinematics (or uses auxiliary sensors such as accelerometers to help track the objects between the camera views [ 5 ]), and a re-identification approach that matches features between people detected across multiple camera views and tracks primarily based on matching identities [ 1 ].…”

Section: Appearance Based Re-identificationmentioning

confidence: 99%

Fusing Appearance and Spatio-Temporal Models for Person Re-Identification and Tracking

2020

View full text Add to dashboard Cite

Knowing who is where is a common task for many computer vision applications. Most of the literature focuses on one of two approaches: determining who a detected person is (appearance-based re-identification) and collating positions into a list, or determining the motion of a person (spatio-temporal-based tracking) and assigning identity labels based on tracks formed. This paper presents a model fusion approach, aiming towards combining both sources of information together in order to increase the accuracy of determining identity classes for detected people using re-ranking. First, a Sequential k-Means re-identification approach is presented, followed by a Kalman filter-based spatio-temporal tracking approach. A linear weighting approach is used to fuse the outputs from these models together, with modification of the weights using a decay function and a rule-based system to reflect the strengths and weaknesses of the models under different conditions. Preliminary experimental results with two different person detection algorithms on an indoor person tracking dataset show that fusing the appearance and spatio-temporal models significantly increases the overall accuracy of the classification operation.

show abstract

Section: Appearance Based Re-identificationmentioning

confidence: 99%