AGA: Attribute-Guided Augmentation

By Mandar Dixit, Roland Kwitt, Marc Niethammer, Nuno Vasconcelos. 2016

105 citations recorded by the source corpus.

We consider the problem of data augmentation, i.e., generating artificial samples to extend a given corpus of training data. Specifically, we propose attributed-guided augmentation (AGA) which learns a mapping that allows to synthesize data such that an attribute of a synthesized sample is at a desired value or strength. This is particularly interesting in situations where little data with no attribute annotation is available for learning, but we have access to a large external corpus of heavily annotated samples. While prior works primarily augment in the space of images, we propose to perform augmentation in feature space instead. We implement our approach as a deep encoder-decoder architecture that learns the synthesis function in an end-to-end manner. We demonstrate the utility of our approach on the problems of (1) one-shot object recognition in a transfer-learning setting where we have no prior knowledge of the new classes, as well as (2) object-based one-shot scene recognition. As external data, we leverage 3D depth and pose information from the SUN RGB-D dataset. Our experiments show that attribute-guided augmentation of high-level CNN features considerably improves one-shot recognition performance on both problems.

Subfields: Aerospace Engineering, Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Low-Shot Learning from Imaginary Data (2018, methodology): “If one has a probability density on transformations, then one can generate additional examples for a novel class by applying sampled transformations to the provided examples [20, 5, 13].”
Data Augmentation in Emotion Classification Using Generative Adversarial Networks (2017, methodology): “As for the existing techniques of data augmentation, they can be grouped into two main types: (a) geometric transformation which is relatively generic and computationally cheap and (b) task-specific or guided-augmentation methods which are able to generate synthetic samples given specific labels [6]”
Few-Shot Unsupervised Image-to-Image Translation (2019, background): “Several recent works propose augmenting the training set for the few-shot classification task by generating new feature vectors corresponding to novel classes [9, 14, 47].”
Feature Transfer Learning for Deep Face Recognition with Under-Represented Data (2018, methodology): “ong regularization [14] or utilizing nonparametric classiﬁcation methods based on distance metric learning [34,39]. Generative model based methods have also been studied in recent years. Dixit et al. [9] propose a data augmentation method using attribute-guided feature descriptor for generation. The”
Global Evolution of Research in Artificial Intelligence in Health and Medicine: A Bibliometric Study (2019, methodology): “In general, AI techniques that are applicable to situations with limited amounts of labeled training data are of great interest for many applications [42], in addition to health and medicine.”
Low-Shot Visual Recognition by Shrinking and Hallucinating Features (2016, background): “ave shown promise in restricted domains such as hand-written characters [28, 25]. They also work well in datasets without much intra-class variation or clutter, such as Caltech 101 [13]. Dixit et al. [9] leverage a corpus with attribute annotations to generate additional examples by varying attribut”
Feature Transfer Learning for Face Recognition With Under-Represented Data (2019, methodology): “[9] propose a data augmentation method using attribute-guided feature descriptor for generation.”
Low-shot visual object recognition (2016, background): “ave shown promise in restricted domains such as hand-written characters [28, 25]. They also work well in datasets without much intra-class variation or clutter, such as Caltech 101 [13]. Dixit et al. [9] leverage a corpus with attribute annotations to generate additional examples by varying attribut”
Augmenting Image Classifiers Using Data Augmentation Generative Adversarial Networks (2018, background): “Various attempts at augmenting features instead of data are investigated in [39, 8].”