Analysing objects in a set of frames

A method of analysing objects in a first frame and a second frame is disclosed. The method includes segmenting 120 the frames to produce masks identifying pixels in potential objects and extracting 130 feature vectors characterising the potential objects. At least one potential object in the first f...

Full description

Saved in:
Bibliographic Details
Main Authors Aria Ahmadi, David Walton, Cagatay Dikici
Format Patent
LanguageEnglish
Published 05.04.2023
Subjects
CALCULATING
COMPUTING
COUNTING
IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:A method of analysing objects in a first frame and a second frame is disclosed. The method includes segmenting 120 the frames to produce masks identifying pixels in potential objects and extracting 130 feature vectors characterising the potential objects. At least one potential object in the first frame is matched 140 with a corresponding potential object in the other frame based, at least in part, on the feature vectors and the object masks. The matching comprises rejecting potential objects based on one or more of an object confidence score, a mask confidence score and a mask area. The motion between the frames of the at least one matched object may be estimated 150. The matching may comprise clustering potential objects in the first and second frames based on the feature vectors and masks. The masks and feature vectors may be generated by a machine learning algorithm. Also provided are processing systems configured to carry out these methods.
Bibliography:Application Number: GB202114845