Automatic Attribute Discovery and Characterization from Noisy Web Data

Abstract

Abstract. It is common to use domain specific terminology -attributes -to describe the visual appearance of objects. In order to scale the use of these describable visual attributes to a large number of categories, especially those not well studied by psychologists or linguists, it will be necessary to find alternative techniques for identifying attribute vocabularies and for learning to recognize attributes without hand labeled training data. We demonstrate that it is possible to accomplish both these tasks automatically by mining text and image data sampled from the Internet. The proposed approach also characterizes attributes according to their visual representation: global or local, and type: color, texture, or shape. This work focuses on discovering attributes and their visual appearance, and is as agnostic as possible about the textual description.

Fine-Grained Visual Comparisons with Local Learning

“…Compared to an existing Shoes attribute dataset [4] with relative attributes [21], UT-Zap50K is about 3.5× larger, offers meta-data and 10× more comparative labels, and most importantly, specifically targets fine-grained tasks.…”

Discovering states and transformations in image collections

“…As a result, there are many great datasets that cover object [9,7,31,25,20], attribute [16,1,10], material [26], and scene categories [32,33]. Here, our goal is to create an extensive dataset for characterizing state variation that occurs within image classes.…”

“…Unlike this work, we investigate discrete, nameable transformations, like crinkling, rather than working in a hard-to-interpret parameter space. Photo collections have also been mined for storylines [15] as well as spatial and temporal trends [18], and systems have been proposed for more general knowledge discovery from big visual data [21], [1], [3]. Our paper differs from all this work in that we focus on physical state transformations, and in addition to discovering states we also study state pairs that define a transformation.…”

Attribute Pivots for Guiding Relevance Feedback in Image Search

“…We validate with three public datasets: Shoes [1], with the attributes from [8] (14,658 images and 10 attributes); outdoor Scenes (2,688 images and 6 attributes); and PubFig celebrity Faces [10] (772 images and 11 attributes). We concatenate GIST and color features for Shoes and Faces, and GIST alone for Scenes.…”