Beyond the sum of parts

• Verfasst von: Yarlagadda, Pradeep Krishna [VerfasserIn]
• Verfasst von: Ommer, Björn [VerfasserIn]
• Titel: Beyond the sum of parts
• Titelzusatz: voting with groups of dependent entities
• Verf.angabe: Pradeep Yarlagadda, Björn Ommer
• Jahr: 2015
• Jahr des Originals: 2014
• Umfang: 14 S.
• Fussnoten: Gesehen am 10.07.2020 ; Date of Publication: 16 October 2014
• Titel Quelle: Enthalten in: Institute of Electrical and Electronics EngineersIEEE transactions on pattern analysis and machine intelligence
• Ort Quelle: New York, NY : IEEE, 1979
• Jahr Quelle: 2015
• Band/Heft Quelle: 37(2015), 6, Seite 1134-1147
• ISSN Quelle: 1939-3539
• DOI: doi:10.1109/TPAMI.2014.2363456
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: category theory
• Sach-SW: category-level object detection
• Sach-SW: cluttered scene
• Sach-SW: computational complexity
• Sach-SW: Computational modeling
• Sach-SW: dense sampling
• Sach-SW: dependent entity
• Sach-SW: detection accuracy
• Sach-SW: Feature extraction
• Sach-SW: grouping
• Sach-SW: Grouping
• Sach-SW: Hough transforms
• Sach-SW: hough voting
• Sach-SW: Hough Voting
• Sach-SW: Hough voting method
• Sach-SW: image representation
• Sach-SW: image sampling
• Sach-SW: joint optimization
• Sach-SW: Joints
• Sach-SW: object detection
• Sach-SW: Object detection
• Sach-SW: object hypotheses
• Sach-SW: object property
• Sach-SW: optimisation
• Sach-SW: popular representation scheme
• Sach-SW: probabilistic Hough voting
• Sach-SW: probability
• Sach-SW: recognition
• Sach-SW: Recognition
• Sach-SW: semi-local image feature
• Sach-SW: sliding window baseline
• Sach-SW: Training
• Sach-SW: Transforms
• Sach-SW: Vectors
• Sach-SW: visual learning
• Sach-SW: Visual learning
• K10plus-PPN: 1724376497

Abstract

The high complexity of multi-scale, category-level object detection in cluttered scenes is efficiently handled by Hough voting methods. However, the main shortcoming of the approach is that mutually dependent local observations are independently casting their votes for intrinsically global object properties such as object scale. Object hypotheses are then assumed to be a mere sum of their part votes. Popular representation schemes are, however, based on a dense sampling of semi-local image features, which are consequently mutually dependent. We take advantage of part dependencies and incorporate them into probabilistic Hough voting by deriving an objective function that connects three intimately related problems: i) grouping mutually dependent parts, ii) solving the correspondence problem conjointly for dependent parts, and iii) finding concerted object hypotheses using extended groups rather than based on local observations alone. Early commitments are avoided by not restricting parts to only a single vote for a locally best correspondence and we learn a weighting of parts during training to reflect their differing relevance for an object. Experiments successfully demonstrate the benefit of incorporating part dependencies through grouping into Hough voting. The joint optimization of groupings, correspondences, and votes not only improves the detection accuracy over standard Hough voting and a sliding window baseline, but it also reduces the computational complexity by significantly decreasing the number of candidate hypotheses.