Object count/area graphs for the evaluation of object detection and segmentation algorithms

By Christian Wolf, J. Jolion. 2006

359 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

Synthetic Data for Text Localisation in Natural Images (2016, methodology): “We report text detection performance using two protocols commonly used in the literature — (1) DetEval [45] popularly used in ICDAR competitions for evaluating localisation methods, and (2) PASCAL VOC style intersection-over-union overlap method ( ≥ 0.5 IoU for a positive detection).”
Real-time scene text localization and recognition (2012, methodology): “Using the ICDAR 2011 competition evaluation scheme [22], the method achieves the recall of 64.”
ICDAR 2013 Robust Reading Competition (2013, background): “Results: Task 1 – Text Localization Performance Evaluation Methodology • Methodology proposed by Wolf and Jolion [1] • Takes into account both • Bounding Box area overlapping and • Precision at the level of detection counts • Possible to create meaningful cumulative results over many images • Ways t”
Evaluation of video activity localizations integrating quality and quantity measurements (2014, methodology): “For this reason, an integral part of the proposed performance evaluation framework is a set of graphs which illustrate this dependence, similar to the graphs proposed in [52].”
A Unified Framework for Multioriented Text Detection and Recognition (2014, methodology): “Moreover, the previous evaluation protocol is replaced by the evaluation method proposed by Wolf et al. [67], as the latter is able to handle the cases of one-to-many and many-to-many matches.”
Symmetry-based text line detection in natural scenes (2015, background): “[38] considers three matching cases: one-to-one, one-tomany and many-to-many.”
Text-Attentional Convolutional Neural Network for Scene Text Detection (2015, methodology): “We follow standard evaluation protocol of the ICDAR 2011 by using the DetEval tool offered by authors of [62], and the ICDAR 2013 standard in [63].”
Detecting Text in Natural Image with Connectionist Text Proposal Network (2016, methodology): “For the ICDAR 2011 we use the standard protocol proposed by [30], the evaluation on the ICDAR 2013 follows the standard in [19].”
Robust Scene Text Detection with Convolution Neural Network Induced MSER Trees (2014, neutral): “[30].”
Scene text detection and recognition: recent advances and future trends (2015, background): “[84] considers three matching cases: one-to-one, one-to-many and many-to-many.”
TextProposals: A text-specific selective search algorithm for word spotting in the wild (2016, methodology): “For the task of text localization the dataset has ground-truth defined at the word level with axis oriented bounding boxes and the proposed evaluation framework is detailed in [57, 61].”
Scene Text Detection and Recognition: The Deep Learning Era (2018, methodology): “one match. It does not consider one-to-many, many-to-many, and many-to-one matchings, which underestimates the actual performance. Therefore, ICDAR2011/2013 follows the method proposed by Wolf et al. [160]. Precision and recall are computed as follows: precision(G;D;t r;t p) = P j Match D(D j;G;t r;”
Text Detection and Recognition in Imagery: A Survey (2015, methodology): “, the number of correct detections and the number of false alarms [50].”
Total-Text: A Comprehensive Dataset for Scene Text Detection and Recognition (2017, methodology): “Like ICDARs datasets [12], Total-Text uses DetEval [23].”
TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes (2018, methodology): “The evaluation protocol is DetEval [35], the same as Total-Text.”
ICDAR 2011 Robust Reading Competition Challenge 2: Reading Text in Scene Images (2011, methodology): “Some problems were reported about the dataset (slightly larger bounding boxes, inconsistent definition of a ”word” for instance whether a hyphen breaks the words or not) as well as the evaluation scheme (handling of one-to-many and manyto-one matches) used in the previous competitions [15].”
Scene Text Localization and Recognition with Oriented Stroke Detection (2013, methodology): “the ICDAR 2011 competition evaluation protocol [18], the method reaches the recall of 66.”
TextBoxes: A Fast Text Detector with a Single Deep Neural Network (2016, neutral): “Results are evaluated under two different evaluation protocols, the DetEval (Wolf and Jolion 2006) and the ICDAR 2013 evaluation (Karatzas et al. 2013).”