Imagine taking a snapshot of a poster with your iPhone and then be able to OCR the text on it. Or scan a book without pressing it hard onto the flatbed glass plate, breaking its back in the process, and still being able to OCR by compensating for the curvature. Or OCR a screenshot… Prizmo 2 promises all that and more, but does it perform?
Prizmo 2 by Creaceed is a software tool that allows you to OCR images that normally aren’t recognised by OCR software. Examples are: skewed images, images on screen that you cannot select with the mouse, photos of posters, paper sheets and business cards, and scans of books with a curvature at the binding.
Prizmo allows you to import these images and perform OCR on them after editing with the built-in tools. Additionally, the Pro version has batch processing features, Automator Actions, and integrates with custom scripts.
I tested Prizmo 2 with a photo of an 800-page book, a scan of that same book, a photo of a POP display and a screenshot of an area inside an application’s interface.
I started with the “Snap Text from Screen” feature. This asks you to drag a selection around a screen area, then to choose from an OCR language. When you click the “Copy Text” button, it will OCR the selection and send the results to the clipboard.
Prizmo’s results were spot-on, literally returning this:
“Ozone 5 Ozone 5 Dynamics Ozone 5 Imager Ozone 5 Maximizer Ozone 5 Equalizer Ozone 5 Reverb Ozone 5 Exciter Mete r Tap Volume Simple Synthesis Normalize”.
I repeated this test using a screenshot of the same area, and the results were again 99% accurate, which is quite astonishing if you realise that I did not crank up the resolution beyond the screen res of 72dpi.
When using Prizmo 2 with screenshots, scans or photos, you will get a complete array of editing tools at your disposal. In many cases, you won’t need them, but it is often possible to improve the OCR results to a full 100% by using some of them.
All tools are readily available in the bottom toolbar. The Crop tool is the most important, as it allows you to focus on exactly what you want to convert to readable text. The Crop feature also holds the tools to adjust the frame, perspective, and curvature. I found that creating an exact crop makes a huge difference. Do it sloppy, and you’ll probably end up with poor OCR results. Take your time, and the score goes up dramatically.
The Crop tool does have an Auto-Adjust button for images that are already near-perfect for OCR.
I found Prizmo 2 to perform brilliantly on photos shot from an angle, even with perspective. However, I also found the curvature feature only to work well when the text isn’t too curved or the curvature doesn’t “run into” the text too much.
To further improve OCR results, Prizmo 2 comes with a unique preview of what the scanned or photographed page looks like to the OCR software. If you can’t make that look more or less clean, then OCR results will be less than ideal. Additionally, Prizmo 2 has a text colour reverse functionality so you can OCR white text on dark backgrounds.
While there’s much more to be told about Prizmo 2, I think it’s clear this app is excellent for capturing live text of anything that you can photograph or see on screen. Next time you see a catch phrase or a stunning marketing paragraph that you would like to edit yourself and adapt to your own needs, use Prizmo to convert it to editable text.
The Pro version has batch processing features, so if you are a student, for example, and you scan book pages in a library with your iPhone, Prizmo Pro will allow you to process all pages based on the first image. The automation features offered by Atomator Actions and Applescript support further increase the Pro version’s power and flexibility, but I haven’t tested those.
Prizmo Pro 2 costs just cents short of 80.00 Euros.
visualsproducer.wordpress.com 