Unsurprisingly, the quality of optical character recognition (OCR) is heavily dependent on the quality of the document photos you feed to the system. Digital workers, or software robots, big data and even machine learning technologies are helpless if the text on the photo is downright illegible.
The good news is that it only takes you a few moments of careful preparation, and then ABBYY's technology will do the work to complete the process. Here are seven mobile scanning tips to overcome common challenges and get the best results scanning from your phone.
1. The more light you have, the better
While photographing documents from your phone with FineReader for iOS, the more light you have, the better. It also helps if the light is evenly distributed.
Taking snapshots at night while sitting in a car with business cards or documents placed in your lap will almost certainly not work. No technology, however intelligent, will be able to capture text that you can hardly read yourself.
2. Point your lens on a high-contrast area to adjust focus
Digital cameras use special technology to determine the distance to the subject. This sometimes means the object in focus may not be the one you want. To perform reliably, the camera needs sufficient lighting and high-contrast objects.
Give your mobile camera a helping hand by pointing the lens to a high-contrast (e.g. black and white) area. It’s best to give the camera some time to focus on the text instead of taking a few hasty snapshots.
3. Hold your breath—or try a flexible clamp mount for sharper images
A blurry photo is more likely to occur in dark environments because the camera will keep the shutter open until the sensor gets enough light. Even if you have a steady hand, any movement to the camera will cause a blur.
If your camera is not equipped with optical image stabilization, shoot like a marksman: hold the camera steady in your hands, hold your breath, and softly press the shutter release button. If you need to take several pictures in a session, we suggest using a flexible clamp mount. We use one ourselves and can confirm that it does the trick.
4. Position your camera to avoid glares and shadows
As we mentioned in our first tip: lighting is everything. Remember, light should be evenly distributed over the document. Using the smartphone flash may help, although bear in mind shadows and glares.
When scanning glossy paper, position the smartphone camera so that you don’t see the direct reflection of the flash on the document. Remember that your arms can obstruct the light and cause shadows, too. If this is the case, you may want to reposition the camera, move to another room, or add an extra light source.
5. Use a solid, dark background
Even when the background is not the subject, your mobile camera may still attempt to pull it into focus. As a general rule of thumb, place your documents against a uniform background and have as few other objects in the viewfinder as possible.
We also suggest avoiding white backgrounds.
6. Position yourself directly above the document for the best angle
If you shoot at an angle, rather than positioning the camera directly above the document, make sure that the angle is not too steep. Otherwise, the image correction algorithms may not help.
While the trapezoid correction mechanism can work wonders, it’s not omnipotent, and won’t be able to restore the document to its rectangular self if you shoot it from a horizontal angle.
7. Inspect your camera specs, not all cameras are created equal
If you plan to use your smartphone as a portable scanner, take a closer look at its camera specs. It should produce a sharp picture across the entire frame, be able to adjust white balance reasonably well, not distort object boundaries, and perform well in different lighting environments.
Smartphones tend to be equipped with better cameras than tablets.
For the best quality scans from your mobile device, follow our tips: use uniform non-white backgrounds, even natural lighting, and steady hands for a picture that’s perfectly suitable for OCR.
Try FineReader for iOS, the first AI-powered scanner to quickly capture docs and books, create copies in PDF and JPEG, and extract texts from scans.