Peekaboom: a game for locating objects in images

By Luis von Ahn, Ruoran Liu, Manuel Blum. 2006

650 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

Training Classifiers with Natural Language Explanations (2018, background): “These explanations are parsed into labeling functions that convert unlabeled data into a large labeled dataset for training a classiﬁer. text (Zaidan and Eisner, 2008; Arora and Nyberg, 2009), or marking relevant regions in images (Ahn et al., 2006).”
Global-and-local attention networks for visual recognition (2018, methodology): “We identiﬁed six mid-to high-level visual layers in ResNet-50 to use with GALA (layers 24, 27, 30, 33, 36, 39; each belonging to the same ResNet-50 processing block).”
GuessWhat?! Visual Object Discovery through Multi-modal Dialogue (2016, methodology): “The first ESP game [4] was developed to collect image tags, and was later extended to Peekaboom [5] to gather object segmentations.”
Mastering the Dungeon: Grounded Language Learning by Mechanical Turker Descent (2017, neutral): “ReferItGame (Kazemzadeh et al., 2014) and Peekaboom (Von Ahn et al., 2006) have similar ideas but for localizing objects.”
Utility data annotation with Amazon Mechanical Turk (2008, background): “This is not the fault of the annotators, but is rather a sign for us to give better instructions.”
Driving in the Matrix: Can virtual worlds replace human-generated annotations for real world tasks? (2016, background): “A variety of approaches have been proposed to efficiently label large amounts of training data including crowdsourcing, gamification, semi-supervised labeling, and Mechanical Turk [3]–[5].”
Crowdsourcing in Computer Vision (2016, neutral): “In spatial annotation games such Peekaboom [von Ahn et al., 2006b], only one worker is shown the image.”
Gaussian Processes for Object Categorization (2010, background): “A number of inge-nious schemes have been developed to obtain labeled data from people performing other tasks (e.g., von Ahn et al. 2006; von Ahn and Dabbish 2004), or directly labeling objects in images (http://labelme.csail.mit.edu/).”
A Survey of Crowdsourcing Systems (2011, methodology): “In addition to image annotation, the Peekaboom system [81] can help determine the location of objects in images, and the Squigl system [2] provides complete outlines of the objects”
Learning visual biases from human imagination (2014, background): “Human-in-the-Loop: The idea to transfer biases from the human mind into object recognition is inspired by many recent works that puts a human in the computer vision loop [6, 27], trains recognition systems with active learning [33], and studies crowdsourcing [34, 31].”
The segmented and annotated IAPR TC-12 benchmark (2010, methodology): “[21] is another game that uses the image annotations generated by the ESP game.”
A Survey on Learning to Hash (2016, methodology): “ism [131] and Caltech 101 [28], GIST features [112] from LabelMe [119] and Peekaboom [140], as well as some features used in object retrieval: Fisher vectors [116] and”
What are the Visual Features Underlying Human Versus Machine Vision? (2017, background): “Closely related to Clicktionary is the ESP game for identifying objects in real-world images [33] and the Peeka-boom game for locating them [34].”
Active Learning with Gaussian Processes for Object Categorization (2007, background): “, [30, 29]), or directly labeling objects in images [1].”