Readspeaker menu

Part Discovery

How to learn distinctive object parts?

Marcel Simon, Alexander Freytag, Erik Rodner, and Joachim Denzler


Motivation

Most approaches in the field of fine-grained recognition rely on human annotated part locations for training. However, semantic parts are not necessarily the best for discrimination. In addition to this, it prevents approaches in this field to be applied to generic datasets like ImageNet. In the area of part discovery, we aim to identify discriminative object parts in an unsupervised way. This will allow us to improve recognition rates in fine-grained tasks and to transfer knownledge from this area to other tasks.


Discovering Part Detectors in CNNs

  Website-Topic-Teaser

[Simon14:PDD]

Marcel Simon and Erik Rodner and Joachim Denzler. Part Detector Discovery in Deep Convolutional Neural Networks. Asian Conference on Computer Vision (ACCV). 2014. accepted for publication [pdf] [bib][code]

[Simon14:PLE]

Marcel Simon and Erik Rodner and Joachim Denzler.  Part Localization by Exploiting Deep Convolutional Networks. ECCV Workshop on Parts and Attributes. 2014. [pdf] [bib][code]

Exemplar-specific Part Discovery


  Exemplar-specific Patch Features for Fine-grained Recognition

[Freytag14:ESP]

Alexander Freytag and Erik Rodner and Trevor Darrell and Joachim Denzler. Exemplar-specific Patch Features for Fine-grained Recognition. German Conference on Pattern Recognition (GCPR). 2014.144--156. (Oral). [pdf] [bib][code]

[Freytag14:BFF]

Alexander Freytag and Erik Rodner and Joachim Denzler. Birds of a Feather Flock Together - Local Learning of Mid-level Representations for Fine-grained Recognition. ECCV Workshop on Parts and Attributes. 2014. [pdf] [bib][code]