Part Detector Discovery in Deep Convolutional Neural Networks
arXiv:1411.3159
Abstract
Current fine-grained classification approaches often rely on a robust localization of object parts to extract localized feature representations suitable for discrimination. However, part localization is a challenging task due to the large variation of appearance and pose. In this paper, we show how pre-trained convolutional neural networks can be used for robust and efficient object part discovery and localization without the necessity to actually train the network on the current dataset. Our approach called "part detector discovery" (PDD) is based on analyzing the gradient maps of the network outputs and finding activation centers spatially related to annotated semantic parts or bounding boxes. This allows us not just to obtain excellent performance on the CUB200-2011 dataset, but in contrast to previous approaches also to perform detection and bird classification jointly without requiring a given bounding box annotation during testing and ground-truth parts during training. The code is available at http://www.inf-cv.uni-jena.de/part_discovery and https://github.com/cvjena/PartDetectorDisovery.
Accepted for publication on Asian Conference on Computer Vision (ACCV) 2014
References in corpus (1)
Cited by in corpus (8)
- ImageNet pre-trained models with batch normalization
- Interpreting CNN Knowledge via an Explanatory Graph
- Harnessing Intrinsic Noise in Memristor Hopfield Neural Networks for Combinatorial Optimization
- Interactively Transferring CNN Patterns for Part Localization
- Growing Interpretable Part Graphs on ConvNets via Multi-Shot Learning
- Examining CNN Representations with respect to Dataset Bias
- Low-rank Bilinear Pooling for Fine-Grained Classification
- Fine-Grained Categorization via CNN-Based Automatic Extraction and Integration of Object-Level and Part-Level Features