Thursday, December 09, 2010

Pleco OCR Review: A Chinese Real Time Dictionary



Pleco is a well crafted Chinese to English and English to Chinese dictionary for iOS devices. The basic free app performs the function of a Chinese dictionary very well. However, the software is unique on the App Store because it is very customisable and allows the purchase of 'Add-ons' to add more features to the dictionary and the sheer number of these in-app purchases that are available is extremely impressive. The most interesting and unique of these features is the Pleco OCR, which allows the use of the iPhone camera to recognise Chinese characters and find related dictionary terms instantly and in real-time.




The OCR module can be found in the 'Add-ons' section of the Pleco app. The module itself costs £8.99 (or £7.99 with the opt-in educational discount) or for $14.99/£12.99.

The OCR (object character recognition) software works by recognising shapes and lines from the camera feed of the iPhone, interpreting Chinese characters from these shapes and then matching this data to the dictionary in Pleco. The performance of Pleco OCR is extremely impressive. The time taken from loading the OCR app to translating a word can take seconds. With other Chinese dictionary applications, it would take considerably longer to input the Chinese character into the software to find the definition. With an application like Pleco OCR, entire strings of characters can be translated on the fly, and the time saving is very considerable.

Tests were performed using an iPhone 3GS. According to Michael Love, the developer, iOS devices without autofocus (even the most recent iPod Touch 4th generation) won't be able to deliver a clear enough image for the software to interpret. This means that currently the only devices that it supports are the iPhone 3GS and the iPhone 4. It's clear that autofocus is really key to using the software, as the OCR relies on having a clear image of the text that one is trying to recognise. As the iPhone is a handheld device, autofocus is constantly being engaged because as the distance between the iPhone camera and the characters constantly changes, causing the image to go out of focus. The OCR provides a large button to autofocus, but the camera seems to autofocus automatically as well. All-in-all, the OCR works quickly, although it can take some time to keep the camera steady enough for the OCR to function.

The 'span lines' function allows the software to recognise a character at the end of one line, and then continue capturing at the beginning f the next line. There is also a history function which allows reviewing of recognised characters. 
The real-time capture can be fairly hectic when trying to translate full sentences, and this feature allows one to review the words that one has scanned.


When using the software, one must make sure that the green bounding box at the centre of the OCR feed was the correct size and that it was translating what one wanted it to translate. By making the box larger than the target character, the OCR would try to interpret the edges of other characters as full characters. Making the box too small would give inaccurate readings. If the box spanned a large number of characters, it would only translate the first character, or a set of characters if it matched a dictionary term. For example, in the above video, I was impressed that it could interpret multi-character phrases such as "來不及".

However, the software has a problem understanding where a word starts and where a word ends. In English, words are clearly differentiated from each other by the spaces between clusters of letters. In Chinese however, a word can be made up of multiple characters and one can only really know a word is separate (and therefore take up a separate entry in the dictionary) by its meaning. When using the green bounding box, one may sometimes miss out a character that makes up the beginning or end part of a word, and therefore miss out on understanding that entire sentence.

Fortunately, this is less a problem with the software itself and is more of a problem with the way the Chinese language is structured as a whole. In truth, Pleco OCR is not best suited for translating full sentences, but performs admirably at being an aid at translating characters in strings of one to three or four characters at a time. The 'history' function is an excellent way of keeping track of large numbers of words. Pleco OCR can be a great aid to quickly translating words as long as one understands the advantages and disadvantages of using it as a main translation device.


Chinese dictionaries as a whole are extremely confusing and time consuming because there are so many different systems for organising the order of characters from the brush stroke count to the character radical. With the advent of touch screen dictionaries such as apps available on the iPhone, this process has become vastly simpler because one is able to input the Chinese character directly into the dictionary. However, this is still a very time consuming process, especially for beginners of the Chinese language who may find difficulty getting accurate results with touch input. Pleco OCR manages to bypass all of these concerns and provides a feature rich, reliable and extremely fast way to translate words. All in all, £8.99 is a very small amount to pay for such a unique, well designed product, and would be an extremely helpful too for anyone interested in learning the Chinese language and wants an extremely quick way to input Chinese characters that they might find in the real world.