Capturing data manually can be a long drawn-out process that can consume a significant amount of time and labour resources. In addition to this, there is always the risk that it may not be 100% accurate, as a result of human error. This last point can be particularly true during periods of heavy workload, or when new staff are being trained. Fortunately, there are digital solutions available today which can not only speed up the process dramatically but also eliminate the majority of these errors.
Optical character recognition: an introduction
Conventional scanning creates a more or less identical digital copy of a paper document, but this digital data is presented as a flat image which is static – it cannot be edited. Optical character recognition (OCR) technology on the other hand is able to actually recognise the actual text that the document contains, enabling it to be quickly and easily edited, searched or compiled.
From a data capture perspective, this means that you can not only significantly reduce expenditure in terms of time and resources, but also dramatically reduce the potential for human error. To do so effectively to secure the highest possible accuracy rate though you need to optimise your processes somewhat. There are a number of key ways in which you can do this.
Tips on how to capture your data with OCR
1. Establish expectations – Are you aiming for a photo-realistic representation of the original with all images and formatting etc. intact, or would a text-only rendering work better for your purposes. OCR software such as Nuance OmniPage can separate the text from its original setting if desirable, as may often be the case for capturing data, but it requires you the user to take this decision first.
2. Plan, organise and communicate the process workflow – Technology is only ever as good as the processes that humans establish for its use. Before you begin capturing data, you should determine each step of the process, and ensure that everyone who is involved in it is aware of the parameters and requirements involved. If one or more individuals are unsure, this can have a negative impact on efficiency and accuracy in the long run. You should also determine which specific file formats should be used at this point, bearing in mind the people who will be required to access the data.
3. Decide on structural elements to include – Will your data include titles, full paragraphs, images, graphs, tables etc? Consider which elements are actually necessary – which are essential, which are ‘nice to have’, and which are unnecessary flab? Streamlining in accordance with these considerations can help you to speed up the process.
4. Get into specifics – If you have control over the documents that will eventually be used in the data capture process (for example, if you are sending out forms to third parties to then fill in), carefully decide which fonts to use for the best clarity, ensure that boxes for handwritten text are big enough to help prevent small illegible writing, and think about any other changes that you can make to help boost the accuracy of the software – ie anything that will improve visibility and clarity.
5. Consider reintroducing the human element – Using optical character recognition removes much of the requirement for human input, but it’s highly recommended that you introduce a watchful human eye to ensure the process is running smoothly. For example, a member of staff could give each document a quick sense check to ensure accuracy, or random human checks could be introduced as quality assurance.
6. Consult the experts – If in doubt, get in touch with your software provider over any issues or questions you may have, as they will be perfectly place to answer your query before it has an impact on your productivity.
If you have further questions about OCR technology take a closer look at Nuance’s OCR document imaging solutions.