Neural Activation Constellations: Unsupervised Part Model Discovery with Convolutional Networks
This addresses the problem of fine-grained and generic object classification for computer vision researchers by providing an unsupervised method that eliminates the need for costly annotations, though it builds incrementally on existing neural network techniques.
The paper tackles unsupervised part model discovery for object recognition by learning constellations of neural activation patterns from convolutional networks, achieving state-of-the-art performance on datasets like CUB200-2011 and Stanford Dogs without part or bounding box annotations.
Part models of object categories are essential for challenging recognition tasks, where differences in categories are subtle and only reflected in appearances of small parts of the object. We present an approach that is able to learn part models in a completely unsupervised manner, without part annotations and even without given bounding boxes during learning. The key idea is to find constellations of neural activation patterns computed using convolutional neural networks. In our experiments, we outperform existing approaches for fine-grained recognition on the CUB200-2011, NA birds, Oxford PETS, and Oxford Flowers dataset in case no part or bounding box annotations are available and achieve state-of-the-art performance for the Stanford Dog dataset. We also show the benefits of neural constellation models as a data augmentation technique for fine-tuning. Furthermore, our paper unites the areas of generic and fine-grained classification, since our approach is suitable for both scenarios. The source code of our method is available online at http://www.inf-cv.uni-jena.de/part_discovery