Automatic Attribute Discovery and Characterization from Noisy Web Data

By Tamara L. Berg, Alexander C. Berg, Jonathan Shih. 2010

499 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Molecular Biology

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Molecular Biology

Citations in the corpus

Describing Textures in the Wild (2013, background): “The study of perceptual attributes of textures has a long history starting from pre-attentive aspects and grouping [16], to coarse high-level attributes [1, 2, 33], to some recent work aimed at discovering such attributes by automatically mining descriptions of images from the Internet [3, 12].”
TreeTalk: Composition and Compression of Trees for Image Descriptions (2014, background): “One challenge in utilizing naturally existing multimodal data, however, is the noisy semantic alignment between images and text (Dodge et al., 2012; Berg et al., 2010).”
PANDA: Pose Aligned Networks for Deep Attribute Modeling (2013, background): “[1] proposed automatic attribute vocabularies discovery by mining unlabeled text and image data sampled from the web.”
Bird Species Categorization Using Pose Normalized Deep Convolutional Nets (2014, methodology): “The features are computed by applying deep convolutional nets to image patches that are located and normalized by the pose.”
Discovering states and transformations in image collections (2015, background): “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].”
Label-Embedding for Image Classification (2015, neutral): “Several works have also considered the problem of discovering a vocabulary of attributes [5], [16], [36].”
Deeply learned face representations are sparse, selective, and robust (2014, background): “a-personal variations [28]. By combining features learned from many face regions, [28] achieved the current state-of-the-art (99:15%) of face veriﬁcation on LFW. Attribute learning is an active topic [10, 2, 25, 4, 24, 40]. There have been works on ﬁrst learning attribute classiﬁers and using attrib”
Learning the Visual Interpretation of Sentences (2013, methodology): “These include object [12], pose [37], facial expression [11], and attribute [9, 4, 26] detectors.”
Learning beautiful (and ugly) attributes (2013, methodology): “These learned attributes have many applications including aesthetic quality prediction, image classification and retrieval.”
Augmented Attribute Representations (2012, background): “Therefore, they required additional human knowledge, obtained either by the analysis of textual sources [16], or by interaction with human users [17].”
Discovering localized attributes for fine-grained recognition (2012, background): “Recent work has introduced visual attributes as intermediate-level features that are both machine-detectable and semantically meaningful (e.g. [2, 3, 6, 9, 11, 14, 28]).”
Multimodal Neural Language Models (2014, methodology): “This dataset was proposed as a means of discovering visual attributes from noisy text [25].”
Start from Scratch: Towards Automatically Identifying, Modeling, and Naming Visual Attributes (2014, methodology): “There are also automatic attribute discovery and naming methods that fall into specific domains like butterfly [37] and fashion productsbengio2009learning recogntion [3].”
How Do You Tell a Blackbird from a Crow? (2013, background): “[2] discover attribute names and classifiers from web data.”
BabyTalk: Understanding and Generating Simple Image Descriptions (2013, background): “Most past approaches have either constructed the set of attribute terms in an ad hoc manner or taken them from an application appropriate ontology [13], [19], [27], [29], but some approaches have tried to learn the attributes directly from image-text pairs [6], [43].”
Mining Motion Atoms and Phrases for Complex Action Recognition (2013, background): “Attribute and part based representations originated from object recognition [4, 22] and have been introduced to action recognition [14, 19, 25].”