Probabilistic multi-word spotting in handwritten text images

Toselli, Alejandro H.; Vidal, Enrique; Puigcerver, Joan; Noya-García, Ernesto

doi:10.1007/s10044-018-0742-z

Cited by 20 publications

(7 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, because most keyword spotting works are focused at a very low level (decide whether or not a word instance in a collection of images is the one that the user was searching for), they disregard this scenario. There are some works that evaluate their systems retrieving paragraphs, with more complex queries [Villegas et al, 2016a, Toselli et al, 2018b].…”

Section: Pagesmentioning

confidence: 99%

A Probabilistic Formulation of Keyword Spotting

Pérez¹

View full text Add to dashboard Cite

Keyword Spotting, applied to handwritten text documents, aims to retrieve the documents, or parts of them, that are relevant for a query, given by the user, within a large collection of documents. The topic has gained a large interest in the last 20 years among Pattern Recognition researchers, as well as digital libraries and archives. vi RESUMEN mos para construir índices de palabras a partir de modelos probabilísticos, basados tanto en un léxico cerrado como abierto. Estos índices son muy similares a los utilizados por los motores de búsqueda tradicionales. Además, se estudia la relación que hay entre la formulación probabilística presentada y otros métodos de gran influencia en el campo de la detección de palabras clave, destacando cuáles son las limitaciones de los segundos. Finalmente, todas la aportaciones se evalúan de forma experimental, no sólo utilizando pruebas académicas estándar, sino también en colecciones con decenas de miles de páginas provenientes de manuscritos históricos. Los resultados muestran que el marco de trabajo presentado permite construir sistemas de detección de palabras clave muy rápidos y precisos, con una sólida base teórica. I com no, als meus pares i la meua germana, que sempre m'han fet costat, m'han animat a fer allò que m'apassionara, i m'han donat estima, tot i el meu caràcter, (a vegades) un tant distant amb ells. També als meus uelos (als quatre), i a la meva tia. En especial al Pepe, que, per fi, deixarà de preguntar-me "encara no has acabat?". Ara ja sí. Allà on vaja, és sempre amb vosaltres.

show abstract

Section: Pagesmentioning

confidence: 99%

A Probabilistic Formulation of Keyword Spotting

Pérez¹

View full text Add to dashboard Cite

show abstract

“…This algorithm enables to train the HMM without any prior segmentation of the training page images into vertical regions. In [28] we can find all the formula to compute the unknown parameters in this case but more specifically for the HTR case (lines and morphemes).…”

Section: The Learning Problem and The Baum-welch Algorithmmentioning

confidence: 99%

“…Foreground classification [5], distinction between print and handwritten text [7], or divide between different photography, printed text and handwritten text [27]. Wei et al [28] also performed a comparison between SVMs, Gaussian Mixture Models (GMMs) and Multi Layer Perceptrons (MLPs) with no clear winner.…”

Section: State Of the Artmentioning

confidence: 99%

“…The calculation of such word probabilistic indexes, was developed as part of approaches to perform word-segmentation-free Key Word Spotting systems [2,14,21,28,29]. With these indexes we now have probabilistic information of the textual contents of the page, without the need to perform a detailed text base line detection, and performing full Handwritten Text Recognition on the detected text lines.…”

Section: Text Content Based Featuresmentioning

confidence: 99%

“…The difference between the detection and extraction outputs is depicted in Fig 1.1. The importance of HTS is due to the need of currently used automatic [1,7,11,25,28,34] or computer assisted [22,26,29] transcription systems to be provided with the specific text line images to be transcribed. The segmentation of text lines and even of the words present in those text lines is also required in most key word spotting methods [8,10,17,23,24] while information indexing methods only require the text lines [2,18,31].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Advances in Document Layout Analysis

Campos¹

View full text Add to dashboard Cite

Handwritten Text Segmentation (HTS) is a task within the Document Layout Analysis field that aims to detect and extract the different page regions of interest found in handwritten documents. HTS remains an active topic, that has gained importance with the years, due to the increasing demand to provide textual access to the myriads of handwritten document collections held by archives and libraries.This thesis considers HTS as a task that must be tackled in two specialized phases: detection and extraction. We see the detection phase fundamentally as a recognition problem that yields the vertical positions of each region of interest as a by-product. The extraction phase consists in calculating the best contour coordinates of the region using the position information provided by the detection phase.Our proposed detection approach allows us to attack both higher level regions: paragraphs, diagrams, etc., and lower level regions like text lines. In the case of text line detection we model the problem to ensure that the system's yielded vertical position approximates the fictitious line that connects the lower part of the grapheme bodies in a text line, commonly known as the baseline.

show abstract