Large-Scale Live Active Learning: Training Object Detectors with Crawled Data and Crowds

By Sudheendra Vijayanarasimhan, Kristen Grauman. 2011

347 citations recorded by the source corpus.

Subfields: 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Global-and-local attention networks for visual recognition (2018, background): “Previous work has shown that it is possible to augment vision systems with human perceptual judgments on difﬁcult problems including face recognition and localization [23–25], bird categorization [26], action recognition [27], and object detection and segmentation [28–31].”
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, background): “Research on using active learning in-the-wild on completely unlabeled data is rare; exceptions are [53, 35].”
Examining CNN representations with respect to Dataset Bias (2017, methodology): “Given a large number of unlabeled samples, existing methods mainly select samples on the decision boundary (Vijayanarasimhan and Grauman 2011) or samples that cannot well fit the model (Long and Hua 2015; Zhang et al.”
Weakly Supervised Object Localization with Multi-Fold Multiple Instance Learning (2015, neutral): “Besides weakly supervised training, mixed fully and weakly supervised [9], active [52], and semi-supervised [40] learning and unsupervised object discovery [11] methods have also been explored to reduce the amount of labeled training data for object detector training.”
Learning Loss for Active Learning (2019, methodology): “For SVMs, distances [52, 53, 27] to the decision boundaries can be used to define uncertainty.”
Adaptive Active Learning for Image Classification (2013, background): “In the realm of computer vision, researchers have adopted active learning in image/video annotation [16, 34, 31], image/video retrieval [29, 12] and image/video recognition [30, 15, 13, 22, 14].”
Y Qualification Main task Y Return N N Y User qualified ? N Feedback Pass ? Pass ? Annotator training Annotating images (2017, background): “Note that the scheme we propose does not require a human-in-the-loop setup [12, 46, 47, 72, 24]: clicks can be acquired separately, independently of the detector training framework used.”
Hierarchical Subquery Evaluation for Active Learning on a Graph (2014, methodology): “Active learning has been successfully applied to many different computer vision problems including tracking [32], image categorization [17], object detection [31], semantic segmentation [30], and image [1] and video segmentation [8], with both human and automatic oracles [18].”
Crowdsourcing Annotations for Visual Object Detection (2012, background): “Active learning (Vijayanarasimhan and Grauman 2011) concerns determining which image to label next (as opposed to a random pick) in a way that better benefits the learning of visual models.”
Leveraging the Wisdom of the Crowd for Fine-Grained Recognition (2016, background): “These tasks yield a great deal of information for a human user and can thus add tremendous value to society.”
Category Modeling from Just a Single Labeling: Use Depth Information to Guide the Learning of 2D Models (2013, background): “[32] proposed semi-supervised learning and active learning to collect objects using an image search engine to sift the raw images.”
Learning Everything about Anything: Webly-Supervised Visual Concept Learning (2014, background): “The work of [49] uses active learning to gather bounding box annotations from Turkers.”
Multi-fold MIL Training for Weakly Supervised Object Localization (2014, neutral): “Besides weakly supervised training, mixed fully and weakly supervised [4], active [33], and semi-supervised [25] learning methods have also been explored to reduce the amount of labeled training data for object detector training.”
NEIL: Extracting Visual Knowledge from Web Data (2013, background): “To minimize human effort, recent works have also focused on active learning [37] [39] which selects label requests that are most informative.”
Flock: Hybrid Crowd-Machine Learning Classifiers (2015, background): “In computer vision, supervised learning approaches have been applied to object classification by asking crowd workers questions to reduce label uncertainty [6], or by highlighting portions of an image corresponding to a given annotation [45].”
We Don’t Need No Bounding-Boxes: Training Object Class Detectors Using Only Human Verification (2016, background): “A few works have proposed active learning schemes specifically for training object class detectors [56, 63].”
What Do I Annotate Next? An Empirical Study of Active Learning for Action Localization (2018, background): “computer vision tasks, such as image classification [4,22,25,36,37,53] and object detection [64], or to construct large-scale image and video datasets [16,66,72].”