Picking Deep Filter Responses for Fine-Grained Image Recognition

Recognizing fine-grained sub-categories such as birds and dogs is extremely challenging due to the highly localized and subtle differences in some specific parts. Most previous works rely on object / part level annotations to build part-based representation, which is demanding in practical applicati...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) s. 1134 - 1142
Hlavní autori: Xiaopeng Zhang, Hongkai Xiong, Wengang Zhou, Weiyao Lin, Qi Tian
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.06.2016
Predmet:
Birds
Detectors
Dogs
Feature extraction
Image recognition
Testing
Training
ISSN:1063-6919
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Recognizing fine-grained sub-categories such as birds and dogs is extremely challenging due to the highly localized and subtle differences in some specific parts. Most previous works rely on object / part level annotations to build part-based representation, which is demanding in practical applications. This paper proposes an automatic fine-grained recognition approach which is free of any object / part annotation at both training and testing stages. Our method explores a unified framework based on two steps of deep filter response picking. The first picking step is to find distinctive filters which respond to specific patterns significantly and consistently, and learn a set of part detectors via iteratively alternating between new positive sample mining and part model retraining. The second picking step is to pool deep filter responses via spatially weighted combination of Fisher Vectors. We conditionally pick deep filter responses to encode them into the final representation, which considers the importance of filter responses themselves. Integrating all these techniques produces a much more powerful framework, and experiments conducted on CUB-200-2011 and Stanford Dogs demonstrate the superiority of our proposed algorithm over the existing methods.
ISSN:1063-6919
DOI:10.1109/CVPR.2016.128