OCR stands for Optical Character Recognition, and it’s exactly what you think it is. Basically, OCR helps in converting hard copies to soft copies. Both text and images, irrespective of the source, can theoretically be converted into machine-encoded text. Users can easily manipulate and edit these files further.
The future is now, and OCR has made life a whole lot simpler. Without further ado, here’s the guide to choosing the best OCR service provider.
Parameters for review
I chose two websites, two Android apps, and one iOS app which were highly recommended. I perused both user ratings (when available) and the opinions of bloggers. Finally, I compiled the following list of software to put to the test.
- i2ocr and Newocr (websites)
- Text Fairy and Google Keep (Android apps)
- CamScanner (iOS app)
I selected four pangrams (sentences containing every letter of the alphabet) and one book cover. Using various backgrounds and text styles, I tried to make them look as diverse as possible.
I uploaded each of these images individually. In addition, I uploaded a page or two of my textbook with plain formatting, just to see how effective each software would be.
Also read: How to block a device from using your WiFi?
Highly recommended by most bloggers, i2OCR is somewhat known for delivering good service for free.
Needless to say, I entered the webpage with certain expectations. Unfortunately, i2OCR was extremely disappointing.
It cannot recognize any text which is not of standard print.
On the book cover, it could not identify even one of the stylized alphabets. The text above the title still has very distinct letters, but i2OCR could not recognize it. It did, however, return “The Original Novels by C. S. Lewis.” This is not a major achievement though, as it is written in plain italics.
i2OCR actually returned completely blank files as the extracted text from the two images with handwritten text! It could not identify even the computer-generated fonts.
I uploaded the collage pictured above as well, only to receive the simple text explainer I’d added as output.
Furthermore, i2OCR is not a secure website. This implies that it is quite dangerous to upload important documents to the site as malicious elements may gain access to it.
This website fell below the mark and I can safely say that I will not be using it anymore.
NewOCR is another well-liked website. One thing I will say is that the website looks sleek and professional. As soon as you open it, you get the impression that they really know what they’re doing.
The first image I uploaded was the handwritten text. I wrote it with my left hand to stimulate a child’s handwriting. NewOCR could not identify the text, even when I highlighted only one word, but some allowances must be made.
The book cover was a huge letdown. All I received as output was long, long strings of gibberish and “The Orzgznal Novels by C. S. Lewis”. They didn’t even spell “original” right!
I uploaded the handwritten text but received nothing in return. Same result with the typed texts. I wondered if it was perhaps my cursive handwriting that the software did not recognize, so I wrote in all lowercase.
I received “fififififi” as the output. Literally. As for the typed text, well…
With fading hope, I tried uploading pages of my Chemistry textbook. Maybe it would read at least plain documents.
To my absolute astonishment, it worked! I actually got a pretty accurate, editable transcript of my Chem textbook.
The best part is, NewOCR has a secure connection. You needn’t worry about uploading sensitive information. Furthermore, you can easily change the orientation of the image and even choose whether to retain column formatting or not.
However, you can upload images only one by one. This isn’t very practical. In addition, I experienced quite a few “504 Gateway Timeout” errors while waiting for my OSR transcripts.
NewOCR may only be suitable for quick, simple document uploads and edits.
Rating: 3/5. Points for the secure connection, for letting users crop, and allowing the change of orientation. Also because it can actually read plain text.
On almost all the websites I checked, Text Fairy was a suggestion. Some loved it, some thought it could do better. I, personally, think Text Fairy deserves points for being forthright.
It didn’t try to oversell itself. When there was a distinct possibility of a scan not working out well, it said so straight off the bat. This saved a lot of time as I didn’t try to scan handwritten text and the blurry typed text to no avail.
Text Fairy couldn’t read the book cover or the archaic text but read the textbook page almost perfectly.
The app managed to read an entire page in a little less than three minutes, with little to no errors! Text Fairy also has the option to instantly translate the text into another language, using a Google Translate plugin. You can also create a PDF on the spot.
Highly efficient, practical, and useful – These are three words that come to mind when I think about this app. I will certainly remember this if and when I need to use OCR tech.
I’ll admit it, I love Google products. I was looking forward to testing this one out. I’d expected it to completely blow me over.
I started by snapping the handwritten ones. The results were not promising, but then again, none of the software had been able to read handwriting.
Next, I tried scanning the book cover. I wasn’t hopeful, but it happened! The dream is alive! Google Keep read not only the text at the bottom but also the title and the text at the top!
Not perfectly, but it was better than the rest. Some alphabets were incorrectly placed, but that can be attributed to the stylized letters. Finally, an app that can extract text from an image as well!
Keep, however, failed at recognizing both the typewritten texts. It did produce a near-exact replica of the textbook. It was formatted in an odd manner, but the text was all there and accurate.
I’d love to give Google Keep a great rating just because it managed to read the book cover. But I couldn’t help but notice how much time it took for each conversion, even if the end product was nonsense. Definitely, it’s easy to use, but I believe the cons outweigh the pros here.
Rating: 3/5. I’m a bit disappointed; I had expected more from a Google product.
Office Lens is apparently a great OCR software. It’s available for both Android and iOS. So is ABBYY FineReader. However, they both require the user to create accounts with the company. I’m aware that not all users would be willing to sign up to use a product of uncertain quality, hence, I’ve chosen to not cover them in this article
I’ll be honest, I was pretty skeptical about CamScanner’s abilities. Last I heard, CamScanner would directly convert your picture into a PDF file and couldn’t do much else.
I was pleasantly surprised when CamScanner made a valiant attempt to decipher my cursive. Half the words were right, half of it was garbage, but I’m not too proud to admit that some of my letters do look similar to what the app thinks they are.
Honestly, I was reeling at the speed of extraction. CamScanner is, by far, the fastest OCR software reviewed here.
I tried the book cover next. To my utter disbelief, CamScanner not only matched Google Keep in being able to detect text but even surpassed Keep! Every character on that page was detected and transcribed.
A little bit disoriented by the dichotomy of CamScanner and most other software on this list, I tried uploading the typed texts.
I had nothing left to say. The competence of CamScanner left me stunned. This app was incredible, and it’s available for both iOS and Android systems.
It was simply too good to be true. Barely daring to hope, I decided to try uploading the text I wrote with my wrong hand. If CamScanner could read this, even a few words, it would undoubtedly be the best app currently on the market. With bated breath, I crop and hit the button to proceed-
– And I’m confronted with this.
That’s right, CamScanner knows it’s wonderful and it knows that you know it too. You can’t avail its OCR services without paying for them first. It doesn’t seem to be too steep a price, especially not with all the offers, but most people don’t have any practical use for it in real life.
Rating: 4.5/5. I’m not upset about the cost, I’m mostly just upset that they didn’t mention it earlier.
In my opinion, websites are the worst and you should never use them. They took too long, your security was at stake, and they just weren’t very good at doing what they’re supposed to.
Pretty much any app can be used to easily transcribe documents. For translational purposes, Text Fairy would be the wisest option. Google Keep can detect text even in bright images, just like CamScanner.
Definitely, the best app of those reviewed today is CamScanner. If you don’t mind paying a bit for incredible product quality, CamScanner is where you should go.
However, for most of us, Text Fairy is the most effective, cost-friendly option of the lot.
Akshaya is an aspiring cardiac surgeon who writes both content and creative pieces at any given juncture. As a powerful orator, a voracious reader, and a bit of a know-it-all, she is usually found with a book in hand. A self-labelled fangirl, Akshaya considers herself a connoisseur of all things nerdy.
Contact Akshaya via email: [email protected] or call: +91-522-4333653