Automatic Evaluation Framework for Word Spotting

Terasawa, Kengo; Imura, Hajime; Tanaka, Yuzuru

doi:10.1109/icdar.2009.119

Cited by 4 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work only relies on closed polygonal lines enclosing areas and makes no further assumption on their geometric structure. Accordingly, straight lines as in [14] are not supported.…”

Section: Performance Evaluationmentioning

confidence: 87%

“…Consequently, the ground truth for the evaluation becomes more complex and contains the location of each word in addition to its transcription. Locations can be expressed as, e. g., lines [14], bounding boxes [11], convex hulls [15], or other polygons. Our work only relies on closed polygonal lines enclosing areas and makes no further assumption on their geometric structure.…”

Section: Performance Evaluationmentioning

confidence: 99%

“…It targets the document retrieval task and judges making hard decisions on document level or on the level of other full entities such as text lines. The automatic evaluation framework proposed by Terasawa et al also incorporates the quality of retrieved word locations by making hard decisions on the required overlap with relevant lines from a ground truth [14]. Rusiñol and Lladós described the complete evaluation process for the symbol spotting task in detail [15].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On Evaluation of Segmentation-Free Word Spotting Approaches without Hard Decisions

Pantke

Märgner

Fingscheidt

2013

2013 12th International Conference on Document Analysis and Recognition

View full text Add to dashboard Cite

Word spotting systems are intended to retrieve occurrences of a given keyword in document images without actually recognizing the full document content. As there is a trend towards segmentation-free word spotting methods, we propose a methodology to evaluate these methods by employing measures that take the quality of the retrieved word locations into account without making hard decisions. We derive a desired evaluation behavior with the help of synthetic examples and show discrepancies of existing evaluation methods. New measures following this behavior are introduced and their differences exemplarily described. The proposed evaluation method is applied to a state-of-the-art word spotting approach.

show abstract

“…Our work only relies on closed polygonal lines enclosing areas and makes no further assumption on their geometric structure. Accordingly, straight lines as in [14] are not supported.…”

Section: Performance Evaluationmentioning

confidence: 87%