Name
Abstract | ||
|---|---|---|
Detecting and correcting skew angles is critical to success in document layout analysis and optical character recognition tasks, as they are more susceptible to failure when on uneven skews. In automation such as postal systems, library management, office business, and banking data entry, skew angle estimation is crucial to improve procedures response. Although different works have addressed this subject, due to the variability in the input data, many solutions are restricted to a specific language, texts whose contents are within a controlled scope, and entries that differentiate printed from handwritten texts. This paper introduces a new method based on RGB gradient capable of detecting and correcting skew angles in different types of documents. We evaluate the proposed method using two public databases and compare our results with other techniques cited in the literature. In general, our proposal achieved results superior to the approaches compared in all groups of documents in the database. Furthermore, we show that our method can work accurately in various text orientations, and it can work efficiently against documents containing short and sparse text lines, non-textual objects, and image noises caused by imperfect scanning. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1007/978-3-031-06430-2_21 | IMAGE ANALYSIS AND PROCESSING, ICIAP 2022, PT II |
Keywords | DocType | Volume |
RGB gradient, Skew detection, Image rotation, Text image processing | Conference | 13232 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 8 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Bruno Rocha | 1 | 0 | 0.34 |
| Gabriel Vieira | 2 | 0 | 0.34 |
| Helio Pedrini | 3 | 0 | 0.34 |
| Afonso Fonseca | 4 | 0 | 0.34 |
| Deborah Fernandes | 5 | 0 | 0.34 |
| Junio Cesar de Lima | 6 | 0 | 0.68 |
| Julio Cesar Ferreira | 7 | 0 | 0.34 |
| Fabrizzio Soares | 8 | 0 | 0.34 |