Back to Publications

Scalable Visual Attribute Extraction through Hidden Layers of a Residual ConvNet

Andres Baloian, Nils Murrugarra-Llerena, Jose M. Saavedra
EventComputer Vision Research Workshop @ CVPR 2021
Research AreasComputational Social Science

Visual attributes play an essential role in real applications based on image retrieval. The traditional manner to build an attribute extractor is by training a convnet-based classifier with a fixed number of classes. However, this approach does not scale for real applications where the number of attributes changes frequently. Therefore in this work, we propose an approach for extracting visual attributes from images, leveraging the learned capability of the hidden layers of a general convolutional network to discriminate among different visual features. We run experiments with a resnet-50 trained on Imagenet, on which we evaluate the output of its different blocks to discriminate between colors and textures. Our results show that the second block of the resnet is appropriate for discriminating colors, while the fourth block can be used for textures. In both cases, the achieved accuracy of attribute classification is superior to 93%. We also show that the proposed embeddings form local structures in the underlying feature space, which makes it possible to apply reduction techniques like UMAP, maintaining high accuracy and widely reducing the size of the feature space.

Download PDF