Frinkiac is an image search engine just for The Simpsons (Wired UK)


It’s never been easier to find the exact moment Mr Burns was shot; Ralph’s valentines’ day card; the moment Bill Gates appeared; or when WIRED was mentioned in The Simpsons. 

Frinkiac is a search engine for Simpsons’ quotes that lets users delve into almost 3,000,000 screenshots from the longest running US cartoon in history.

The screenshots used in ‘Frinkiac’ come from the first 15 episodes of the Simpsons, and can have the quote from the relevant part of the show imposed on top of them. Simpsons’ fans should be warned that many hours may be lost to Frinkiac, which has been created by Paul Kehrer, Sean Schulte, and Allie Young. 

However, those wanting to test out the Frinkiac will probably have to be quick. Fox, who own and produce the Simpsons, are protective of their content; in 2010 the company pulled a Simpson’s intro allegedly created by Banksy, only reinstating it after a backlash.

Young told WIRED the idea for the search engine came about after the other creators were constantly quoting the cartoon at each other. “It eventually led to the idea of building a search engine for finding the exact scene for every quote. At the time, they laughed it off as an outlandishly difficult project and that was that. About 6 months ago it occurred to them that maybe it wasn’t so outlandishly difficult after all and they decided to go for it.” 

The search system works based on an index created from the words of each subtitle — these are broken into prefixes from the subtitles. 

“The more you type, the more accurate your search becomes. For example, ‘kill anyone who looks’ isn’t accurate enough to find Rex Banner, so keep typing, ‘kill anyone who looks at me cockeyed’ to find the scene,” the creators said in a blogpost.

The trio parsed episodes from the 15 series and generates screen captures, they say the coding behind the search engine “attempts to determine” the relevant screenshots to grab in a “fairly naïve way”.

Each scene of all the episodes were cut into 100 equal-sized buckets, the average colour is then taken from the bucket and the colour is compared to the previous bucket. If the average colours were different the frame was saved, data from subtitle files is added and the dataset — all 161GB of it — was then uploaded to a separate server.

“It took a few weeks to write the basic structure of the app,” Young told WIRED. “The parser is reasonably fast — it got through seasons 1-15 and indexed them against the subtitles overnight, which was a good thing because we had to run it a few times to get everything right. As with any app the longest part was making it work nicely (which is where I came in) and that was a few months of basically using it constantly and making it better.” 

The website isn’t perfect yet though, which Young acknowledges and says there are plans for improving the user experience. “We’ve noticed that it has some usability challenges on mobile devices and we’d like to enhance the search engine to support different ways of searching with things like ‘exact string in quotes’ searches, homophones (‘breaks’ vs ‘brakes’), apostrophes, handle weird caption spelling (there are a lot of ways they spelled ‘glaven’ and searchers shouldn’t need to know that).” 


If the article suppose to have a video or a photo gallery and it does not appear on your screen, please Click Here

3 February 2016 | 2:38 pm – Source:


Leave a Reply

Your email address will not be published.