An Intelligent Hybrid Model for Identity Document Classification

Abstract

Digitization, i.e., the process of converting information into a digital format, may provide various opportunities (e.g., increase in productivity, disaster recovery, and environmentally friendly solutions) and challenges for businesses. In this context, one of the main challenges would be to accurately classify numerous scanned documents uploaded every day by customers as usual business processes. For example, processes in banking (e.g., applying for loans) or the Government Registry of BDM (Births, Deaths, and Marriages) applications may involve uploading several documents such as a driver's license and passport. There are not many studies available to address the challenge as an application of image classification. Although some studies are available which used various methods, a more accurate model is still required. The current study has proposed a robust fusion model to define the type of identity documents accurately. The proposed approach is based on two different methods in which images are classified based on their visual features and text features. A novel model based on statistics and regression has been proposed to calculate the confidence level for the feature-based classifier. A fuzzy-mean fusion model has been proposed to combine the classifier results based on their confidence score. The proposed approach has been implemented using Python and experimentally validated on synthetic and real-world datasets. The performance of the proposed model is evaluated using the Receiver Operating Characteristic (ROC) curve analysis.

Enhancing Identity Document Classification in KYC Processes: An Evaluation of the Bag-of-Visual-Words Model and Segmentation Impact

“…Once a matching model was found, a more intricate analysis was carried out to determine whether the matched model should be accepted or rejected. Khandan [10] proposed a method that combined SIFT and OCR for identity document classification. The study aimed to develop a method to classify identity documents with a confidence level for each match, which was determined by the number of matches in the SIFT model during each classification task.…”

“…Some studies have also shown that incorporating visual features can enhance text-based document classification [22]. In the context of identity document classification, visual features often serve as the basis for classification methods [5,6], although some studies explore the combination of multiple types of information such as visual features and textual information [10], or visual features and spatial information [5,9]. CNN has been proposed as a classification method for all three categories examined in our study: general image classification [17], text document classification [19], and identity document classification [11].…”

“…Regarding datasets, we discovered that the scarcity of research on identity documents compared to other document types can be attributed to the sensitive nature of these documents, which often contain confidential information. In certain studies, researchers addressed this challenge by collecting their own data [6,9] or collaborating with companies [10,11] to obtain datasets. However, this approach restricts access to the dataset by other researchers, limiting reproducibility.…”

“…For instance, many identity documents share the same layout regardless of their type or issuing country [6]. Visual features can also be used along with textual features; this approach is demonstrated using SIFT to detect visual features and OCR to extract textual information [10]. However, OCR requires prior knowledge of the expected language within the document, as highlighted in the study.…”