Analysis of barcodes, verification of passports, validation of legal documents - these are just a few examples that can be found in the attachments of e-mails. These use cases of image analysis are part of our everyday life and become more and more numerous. Our IDA Suite for intelligent document processing uses OCR (optical character recognition), which is particularly useful for processing attachments of incoming emails. To understand how can we use it, we interviewed our new partner PLANET AI.
An interview with Jesper Kleinjohann Chief Operating Officer at PLANET AI conducted by Thomas Leduc Marketing Manager at Golem.ai
What is OCR and what are the use cases ?
OCR nowadays is a synonym for capturing text from scanned documents and images meaning raw pixel data. This is a crucial part in every document based process, whether the input channels are e-mails, apps or paper being scanned. The quality of the reading process infects all subsequent tasks in the document process automation.
What are the technological challenges ?
Classical OCR approaches rely on pre-processing steps like binarization and character segmentation to read each and every character. This has strong limitations if the scan or image quality drops or especially for handwritten texts. Our approach processes text as a sequence like we humans do, avoiding all the negative downsides of classical OCR approaches. We preserve all the information offering a deep understanding of the captured text for subsequent tasks.
Why is attachment analysis critical to automating message processing ?
The actual mail body often only contains the cover letter. The majority of the necessary information to implement a full process automation is hidden in the attachment. Those documents are scanned images or pictures taken with a smartphone. Making those information accessible in the best possible quality is the key for any further process automation.
Jesper Kleinjohann
Chief Operating Officer at PLANET AI
PLANET AI Founded in 2015 as a research-driven company, PLANET AI develops software products for text, speech and image recognition.
