Multi-Level Active Prediction of Useful Image Annotations for Recognition

By Sudheendra Vijayanarasimhan, Kristen Grauman. 2008

86 citations recorded by the source corpus.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop (2015, methodology): “In many ways, this process is similar to active learning [14, 18, 3, 20].”
We Don’t Need No Bounding-Boxes: Training Object Class Detectors Using Only Human Verification (2016, background): “Previous active learning work has mainly focussed on image classification [25, 26, 27, 39], and freeform region labelling [45, 54, 55].”
Perceptual Annotation: Measuring Human Vision to Improve Computer Vision (2014, background): “Active learning [1] is a prominent example of such an approach, wherein training set quality is enhanced by placing human annotators “in-the-loop” with a machine learning system [7], [8], [9], [10], [11], [12], [13].”
LabelMe: Online Image Annotation and Applications (2010, background): “Another work for optimizing label propagat ion is [77], where a learner is trained to balance the relative co sts for obtaining different levels of annotation detail, along with the reduction of uncertainty the annotation provides to the system.”
Active Learning by Labeling Features (2009, methodology): “When mixing queries, it will be important to use different costs for different annotation types (Vijayanarasimhan and Grauman, 2008), and estimate the probability of obtaining a useful response to a query.”
A scalable active framework for region annotation in 3D shape collections (2016, background): “Their main advantage is that they acquire training data incrementally, enabling a classifier to predict which samples need to be labeled next to improve their performance [Vijayanarasimhan and Grauman 2008; Branson et al. 2011; Vezhnevets et al. 2012; Branson et al. 2014].”
Large-Scale Live Active Learning: Training Object Detectors with Crawled Data and Crowds (2011, background): “Active learning has been shown to better focus annotation effort for image recognition tasks [6, 7, 9] and region labeling [8, 10].”
From Theories to Queries (2011, background): “Vijayanarasimhan and Grauman (2009a) demonstrated potential cost savings with active learning using predicted annotation costs for computer vision.”
A General Active-Learning Framework for On-Road Vehicle Recognition and Tracking (2010, background): “approach that is taken in many object-recognition studies [8], [15], [29], [34].”
Active Learning Literature Survey (2009, background): “Vijayanarasimhan and Grauman (2009) have extended the idea to SVMs for the image retrieval task, and also explore an approach that interleaves queries at varying levels of granularity.”
Annotator rationales for visual recognition (2011, background): “While recent work explores various issues in collecting useful labeled datasets [1, 4, 12, 19, 21, 24], we are the first to propose asking annotators for explanations of their labels to directly improve visual category learning.”
Best of both worlds: Human-machine collaboration for object annotation (2015, background): “Works such as [26, 13, 54] use multiple modalities of human feedback, with varying costs, and accurately model and predict the success of each modality.”