Part Detector Discovery in Deep Convolutional Neural Networks
This work addresses the problem of part localization for fine-grained classification, offering a more efficient approach that reduces annotation requirements, though it is incremental as it builds on existing pre-trained networks.
The paper tackles the challenge of part localization in fine-grained classification by introducing a method that uses pre-trained CNNs for object part discovery without additional training, achieving excellent performance on the CUB200-2011 dataset and enabling joint detection and classification without bounding box annotations during testing.
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.