Compliance with the Electronic Books Preservation Act and the advancement of digital transformation (DX) have increased opportunities to digitize documents of various sizes, such as forms, receipts, and invoices. However, these documents are often irregular in size, causing issues in scanned images, such as unwanted margins and skew when documents of different sizes are scanned together (Fig. 1).
With previous models, irregular-sized documents had to be scanned as one of the standard sizes, followed by a separate process of cropping the image using editing software or similar tools. This made document digitization time-consuming and labor-intensive.
To address this, we have developed a technology for automatically cropping scanned images of irregular-sized documents during scanning. This technology automatically generates scanned images with reduced unwanted margins and skew, thereby reducing the post-processing work that was previously required. As a result, it streamlines the digitization of documents such as receipts and invoices, helping reduce customers' workload.
The irregular-sized document cropping technology consists of three processes: 1. contrast generation, 2. document area detection, and 3. image correction (Fig. 2). These are implemented through a combination of an image-processing ASIC and image-processing software, enabling image cropping and skew correction to be performed in real time. Specifically, for the document area detection process, we developed a new algorithm to identify the document area. By performing area analysis as soon as the leading edge of the document is read, image correction can be carried out without interrupting the scanning process.
As a result, documents of various sizes can be efficiently digitized while maintaining high scanning productivity, and ready-to-use image data can be obtained without additional editing.