Data‐driven nonrigid object feature analysis: Subspace application of incidence structure
Abstract In object recognition, feature extraction algorithms are designed to capture the discriminate statistics of objects. Due to pose, deformation and background clutter, the recognition of objects becomes nontrivial, particularly nonrigid samples. Through incidence and geometric structure, this...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2020-04-01
|
Series: | Engineering Reports |
Subjects: | |
Online Access: | https://doi.org/10.1002/eng2.12141 |
id |
doaj-51ce8600a9d14b18a6d5ce63c24dcf8d |
---|---|
record_format |
Article |
spelling |
doaj-51ce8600a9d14b18a6d5ce63c24dcf8d2020-11-25T03:01:47ZengWileyEngineering Reports2577-81962020-04-0124n/an/a10.1002/eng2.12141Data‐driven nonrigid object feature analysis: Subspace application of incidence structureNicholas Wells0Chung W. See1Cerebrum Matter Ltd Loughborough UKFaculty of Engineering University of Nottingham Nottingham UKAbstract In object recognition, feature extraction algorithms are designed to capture the discriminate statistics of objects. Due to pose, deformation and background clutter, the recognition of objects becomes nontrivial, particularly nonrigid samples. Through incidence and geometric structure, this article reports on the data‐driven identification of critical features located on object exemplar profiles. The investigation is demonstrated using the features of a cat's head and the application of the Hough transform to extract planar geometric features. A data‐driven recognition routine is described that accumulates prior knowledge for evaluating the error contribution of critical features impacting recognition confidence. In measuring affine facial features, feature parallelism is tracked to determine rotations and elevations of a cat's head. A preliminary recognition error of 8.2% and 17.8% is determined for a front training profile. An analysis proceeds to determine contributions to this error due the identified critical features.https://doi.org/10.1002/eng2.12141data‐driven recognitionfeature analysislinear geometrynonrigid objects |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Nicholas Wells Chung W. See |
spellingShingle |
Nicholas Wells Chung W. See Data‐driven nonrigid object feature analysis: Subspace application of incidence structure Engineering Reports data‐driven recognition feature analysis linear geometry nonrigid objects |
author_facet |
Nicholas Wells Chung W. See |
author_sort |
Nicholas Wells |
title |
Data‐driven nonrigid object feature analysis: Subspace application of incidence structure |
title_short |
Data‐driven nonrigid object feature analysis: Subspace application of incidence structure |
title_full |
Data‐driven nonrigid object feature analysis: Subspace application of incidence structure |
title_fullStr |
Data‐driven nonrigid object feature analysis: Subspace application of incidence structure |
title_full_unstemmed |
Data‐driven nonrigid object feature analysis: Subspace application of incidence structure |
title_sort |
data‐driven nonrigid object feature analysis: subspace application of incidence structure |
publisher |
Wiley |
series |
Engineering Reports |
issn |
2577-8196 |
publishDate |
2020-04-01 |
description |
Abstract In object recognition, feature extraction algorithms are designed to capture the discriminate statistics of objects. Due to pose, deformation and background clutter, the recognition of objects becomes nontrivial, particularly nonrigid samples. Through incidence and geometric structure, this article reports on the data‐driven identification of critical features located on object exemplar profiles. The investigation is demonstrated using the features of a cat's head and the application of the Hough transform to extract planar geometric features. A data‐driven recognition routine is described that accumulates prior knowledge for evaluating the error contribution of critical features impacting recognition confidence. In measuring affine facial features, feature parallelism is tracked to determine rotations and elevations of a cat's head. A preliminary recognition error of 8.2% and 17.8% is determined for a front training profile. An analysis proceeds to determine contributions to this error due the identified critical features. |
topic |
data‐driven recognition feature analysis linear geometry nonrigid objects |
url |
https://doi.org/10.1002/eng2.12141 |
work_keys_str_mv |
AT nicholaswells datadrivennonrigidobjectfeatureanalysissubspaceapplicationofincidencestructure AT chungwsee datadrivennonrigidobjectfeatureanalysissubspaceapplicationofincidencestructure |
_version_ |
1724692061293117440 |