Prizmo 2: OCR Done With Style.

With the advent of the shift to a Paperless world, OCR has gained even greater importance. It’s the often overlooked detail that you may take for granted. That is of course until you try and search a PDF you scanned and realize it’s just an embedded image. Chances are that you may already have a good document scanner that does OCR. If, however, you don’t, then Prizmo 2 may just be what the doctor ordered.

Now I could just say that Prizmo does OCR with style, but the truth is that it does so much more. Read on after the break to find out what I mean.

Like the article? You should subscribe and follow us on twitter.

Interface

Prizmo is a beautifully designed app. The developers have paid great attention to detail and that reflects in every aspect of interaction with the app. All the app’s tools are accessed by means of popovers with smooth and discreet animations, making the app a pleasure to use.

Although Prizmo is rather intuitive and easy to use, clicking the small ? available on many of the popovers shows some help and additional information pertaining to that particular feature. This makes learning to use the app a breeze and something that happens quite naturally.

It’s not all good though. I found the lack of an ok or apply on the popovers unintuitive. In order to accept changes made you simply click the tool button a second time or alternatively close the popover.

Capture

Prizmo really goes to town on capture. I won’t bore you with details about the usual suspects, drag and drop, Open Image File and Browse Image Library. Instead I’ll focus on the various aspects that I think make Prizmo shine in the capture department.

The least obvious, and one I initially dismissed as being a sales gimmick without much use, is the ability to OCR a section of your screen. Clicking the menubar icon immediately dims your display and allows you to select a portion of your screen, choose the desired language and finally copy the OCR’d text. This may seem pointless at first, but I assure you it’s not. Ever tried to copy a description from an app on the Mac App Store? Or maybe some text from a remote machine when the VNC client doesn’t support the clipboard? Now you can!

Moving right along, we come to my favourite method. Capture from connect devices. Selecting Import From Camera or Scanner slides down a panel showing all your connected devices from which you can import files. It even has support for Wi-Fi connect scanners. If you have an iOS device connected and take a photo, Prizmo detects this and immediately imports that photo. Another great feature that worked flawlessly with my testing was capture from a tethered camera. If you have a modern Canon or Nikon DSLR connected, you can easily capture images from within Prizmo with the click of a button.

Prepare Your Image

Depending on your capture method, you will need to do more or less pre-processing work in order to achieve good readable text for OCR. If you’ve used either a document or flatbed scanner, then chances are you won’t need any extra steps and can immediately press Recognise.

Prizmo is smart enough to detect if your image is upside down and automatically rotates it for you.

There are however many instances when you will need to perform some pre-processing before Prizmo can OCR your image. Thankfully along the bottom of the window you’ll find all you need to achieve the best possible end results each time.

The Tools.

• Crop – You can use the crop tool to either select just a portion of the image to OCR, or alternatively, if you’ve captured the image with a camera or smartphone you can use the crop tool to select the document, leaving out the background. There is a button that will try and auto-detect the page borders but when that doesn’t work you can simply drag the guides to match up with the page borders. The crop tool doubles also as a straightening tool allowing you to correct perspective and even compensate for the curvature of open books.

• Dimensions – This allows you to define the size of your capture and to what size or page format it should be adjusted to. If Prizmo detects that the image resolution is not optimal for OCR you will be presented with an option to automatically resize the source for optimal results.
• Lens Correction – Lens correction does just that, it allows you to compensate for any lens distortion. If Prizmo has no calibration profile for your device you can easily create one following the calibration wizard and then saving it for later use.
• Adjust – This seemingly humble popover hides all you will need to perform color corrections such as brightness, contrast and saturation. You can adjust the sharpness and normalise the image choosing the white point and black point. This is fundamental to achieve a good enough contrast so that Prizmo may easily recognise the text on your image.
• OCR – Here you can finally get a feel for what Prizmo will see. You can bump up the resolution and increase readibility until the text becomes clear.
• Language – Finally but not least important is the language selection. Although Prizmo has support for over 40 Languages, you can fine tune the active ones so as not to clutter this popover with unneeded languages.

The File Types

There are essentially three types of files that you can process with Prizmo. Depending on what type of file you’ve chosen, the tools available and the end result will vary somewhat.

1. Text – All the previously mentioned tools are available for this document type. Prizmo will perform normal OCR and allow you to export to various online services such as Evernote, Box, Dropbox, Google Drive or any other WebDav location.
2. Business Card – As with the previous case, all the tools are active and at your disposal. Prizmo will perform OCR and try to correctly identify the various fields after which it will create a new contact sheet with the processed data.
3. Image – For image type document, the app will hide any tool related to OCR. Furthermore the Recognise button will be replaced with the option to Export, usually only found after having OCR’d and image.

Process Your Image

Once your image has been properly prepared it is finally time to process the text. Clicking Recognise, Prizmo will start to perform Optical Character Recognition. Once complete you are presented with a whole new range of tools with which to fine tune the resulting text. Prizmo will identify 3 types of Regions: Text, Numeric, and Image.

Any change you perform on a region automatically updates the recognised text. Add, Delete, Reorder and Merge various regions to suit your needs. If you find that some text was not properly recognised simply adjusting the readability slider on the OCR popover may produce better results. You can also change the recognised text’s language.

You can translate the active region’s text and from my limited testing it functioned pretty well forming legible and cohesive sentences (at least for the languages I could understand and read). Prizmo also has text-to-speech, allowing you to hear out loud the recognised text.

Once you’re finally satisfied with your text you can export your results to various image formats, PDF, PDF (Image + Searchable Text), Rich Text and Plain Text. For the Business Card type, you can save the result as a new Contact, email or save the vCard file. For the Image Type, export options are obviously limited to images and PDF

The Pro Pack

While I’m not much of a proponent for in-app purchases, I feel that the developers found a good compromise in this case. Instead of pricing the app extremely high, full of bells and whistles that the casual user will have little need for, they provide a Pro Pack as an in-app purchase.

The Pro Pack adds Batch Processing of multiple similar documents, an Automator Action and the ability to integrate Prizmo in more complex enterprise workflows through the use of AppleScripts or shell scripts. The Automator Action is particularly useful for those attempting a Paperless workflow. Simply setup an automator workflow to watch a folder, drop in an image and Prizmo does its magic.

Conclusion

Given its array of features, even those with a dedicated OCR app can benefit from Prizmo. From the simple menubar tool to capture difficult text to the translation and text-to-speech, Prizmo is sure to have something to fit your needs. Those requiring the power and flexibility of automation are not left out in the cold while those with more humble needs are catered to with great elegance.

The attention to detail mentioned earlier goes to the point of the developers providing the user guide in the form of an interactive iBook that is a pleasure to read on the iPad. There’s little more you could ask from a modern Mac OCR app, and Prizmo is an app we highly recommend if you want to do advanced OCR on your Mac.