Use case visual Bag-of-Words techniques for camera based identity document classification

Heras, Lluís-Pere de las; Terrades, Oriol Ramos; Lladós, Josep; Fernandez-Mota, David; Cañero, C.

doi:10.1109/icdar.2015.7333856

Cited by 13 publications

(14 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Research teams are forced to create and maintain their own datasets using their own resources. This can be easily observed by looking through the recent papers devoted to identity document analysis, for instance [16,2,24]. For some teams having access to such sensitive data is not a problem, especially if they are industry-oriented, already have a product related to identity document analysis, have government support and possess the required expertise in data security.…”

Section: Introductionmentioning

confidence: 99%

MIDV-500: a dataset for identity document analysis and recognition on mobile devices in video stream

Arlazarov

Bulatov

Chernov³

2019

Computer Optics

View full text Add to dashboard Cite

A lot of research has been devoted to identity documents analysis and recognition on mobile devices. However, no publicly available datasets designed for this particular problem currently exist. There are a few datasets which are useful for associated subtasks but in order to facilitate a more comprehensive scientific and technical approach to identity documents recognition more specialized datasets are required. In this paper we present a Mobile Identity Document Video dataset (MIDV-500) consisting of 500 video clips for 50 different identity document types with ground truth which allows to perform research in a wide This work is partially supported by Russian Foundation for Basic Research (projects 17-29-03170 and 17-29-03370). Source images for MIDV-500 datasets are obtained from Wikimedia Commons

show abstract

Section: Introductionmentioning

confidence: 99%

MIDV-500: a dataset for identity document analysis and recognition on mobile devices in video stream

Arlazarov

Bulatov

Chernov³

2019

Computer Optics

View full text Add to dashboard Cite

show abstract

“…Local feature points, such as SIFT [21], are widely used as local features due to their description capabilities. Regarding the second step, image encoding, BOW were originally used to encode the feature point's distribution in a global image representation [12,16]. Fisher vectors and VLAD later showed improvement over the BOW [23,14].…”

Section: Previous Workmentioning

confidence: 99%

Identity Documents Classification as an Image Classification Problem

Sicre¹,

Awal

Furon³

2017

Image Analysis and Processing - ICIAP 2017

View full text Add to dashboard Cite

To cite this version:Ronan Abstract. This paper studies the classification of identification documents, which is a critical issue in various security contexts. We address this challenge as an application of image classification, a problematic that received a large attention from the scientific community. Several methods are evaluated and we report results allowing a better understanding of the specificity of identification documents. We are especially interested in deep learning approaches, showing good transfer capabilities and high performances.

show abstract

“…Moreover, in the last decade, the performance of mobile devices has been improved every year, allowing them to be able to capture an image, processing it and get results similar to those of a flatbed scanner (Chabchoub, Kessentini, Kanoun, Eglin, & Lebourgeois, 2016;de las Heras, Terrades, Llados, Fernandez-Mota, & Canero, 2015;Kaur & Garg, 2015;Sharma & Sharma, 2016). Features such as the size, mobility and ability to instantly capture images in an easy way, compared to a scanner, have made smartphones receive more attention in their use as a mobile scanner (Bai, Yin, & Liu, 2013;Jacobs, Simard, Viola, & Rinker, 2005;Sharma & Sharma, 2016).…”

Section: Research Contextmentioning

confidence: 99%

“…Usually, pictures taken by smartphone cameras are captured in natural scenes. Distortion of perspective, different styles of text and light sources that introduce shadows and reflections are some of the drawbacks present in the images captured in natural scenes that hinder the process of text recognition (Bai et al, 2013;de las Heras et al, 2015;Zhu & Zanibbi, 2016). Therefore, in order to efficiently segment the text of the rest of the objects, an integration of different DIP algorithms must be performed.…”

Section: Research Contextmentioning

confidence: 99%

“…In the literature, there are some systems able to extract the information already classified based on the template of the identity documents processed (de las Heras et al, 2015;Ryan & Hanafiah, 2015;Simon et al, 2015). The use of templates to extract the textual information from IDs guarantees a fast processing since the location of this one is previously known.…”

Section: Research Contextmentioning

confidence: 99%

See 1 more Smart Citation

End-to-end system for extracting and interpreting testual information of interest from identify documents images.

Menéndez¹,

Carlos²

View full text Add to dashboard Cite

Identity documents (ID) are one of the primary sources for obtaining information about a citizen. The center of many applications within the administrative and service sectors is the extraction of the data contained in ID cards. Therefore, in this research is proposed the implementation of an automated system able to extract and interpret the textual information from identity documents images. The proposed end-to-end system allows the automation of a registration or verification process that requires the acquisition of information about a citizen using his identity documents. The system obtained through this research is considered as an end-to-end system since it covers every stage of the information of interest extraction process from IDs images.Different to the template-based systems, the proposed system uses a semantic attribution algorithm that allows to classify and attribute meaning to the information from IDs according to its semantics. This research is the first comprehensive description of a complete information extraction system to process IDs that describes from image processing to named entity recognition.To evaluate the performance of the research were proposed different metrics based on the internal functions of the system. The final evaluation shows satisfactory results showing that the end-to-end system is capable of extracting and interpreting textual information from identity documents images without prior knowledge of their layouts.

show abstract

Use case visual Bag-of-Words techniques for camera based identity document classification

Cited by 13 publications

References 12 publications

MIDV-500: a dataset for identity document analysis and recognition on mobile devices in video stream

MIDV-500: a dataset for identity document analysis and recognition on mobile devices in video stream

Identity Documents Classification as an Image Classification Problem

End-to-end system for extracting and interpreting testual information of interest from identify documents images.

Contact Info

Product

Resources

About