Predicting Sufficient Annotation Strength for Interactive Foreground Segmentation

By Suyog Dutt Jain, Kristen Grauman. 2013

81 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

Extreme Clicking for Efficient Object Annotation (2017, background): “Papers reporting faster timings [32, 54] aim for lower-quality boxes (e.”
What's the Point: Semantic Segmentation with Point Supervision (2015, methodology): “Providing an accurate outline of a single object takes between 54 seconds [13] and 79 seconds [5].”
PhraseClick: Toward Achieving Flexible Interactive Segmentation by Phrase and Click (2020, background): “Many interactive segmentation methods have been proposed in the past decades using many different interaction types such as bounding boxes [16, 29, 49, 57], contours [1,6, 26,45], strokes [5,31,51,53] and clicks [25,28,34,40,41,58].”
Y Qualification Main task Y Return N N Y User qualified ? N Feedback Pass ? Pass ? Annotator training Annotating images (2017, background): “Papers reporting faster timings [32, 54] aim for lower-quality boxes (e.”
Interactive Object Segmentation With Inside-Outside Guidance (2020, background): “[54] reported 10.21 s while [19] reported 7.0 s ).”
Best of both worlds: Human-machine collaboration for object annotation (2015, methodology): “[6, 26] One key component, in contrast to prior work, is the incorporation of feedback from multiple types of human input and from multiple computer vision models.”
Crowdsourcing in Computer Vision (2016, unknown): “[2012] and Jain and Grauman [2016] introduce active image segmentation systems for semantically segmenting a large set of related images.”
Embracing Error to Enable Rapid Crowdsourcing (2016, methodology): “Other methods vary both the labeling cost and annotation accuracy to calculate a tradeoff between the two [23, 14].”
Click Carving: Segmenting Objects in Video with Point Clicks (2016, methodology): “Timing data from (Jain and Grauman 2013; Lin et al. 2014) indicate it typically takes 54-79 seconds to manually outline an object; we use the more optimistic 54 seconds for this baseline.”