If you’re reading this, you’ve probably already mastered the skill of counting. But just because something is easy doesn’t mean it’s enjoyable, so the makers of a documenting-scanning app for smartphones has added a new AI-powered feature that can automatically count up similar objects in a photo for you. This is truly a gift for anyone cursed with counting inventory.
The iScanner app (which shows up as “PDF Scanner App – Scan Documents with iScanner” in the Google Play Store and “Scanner App: PDF Document Scan” in Apple’s App Store) is yet another example of cleverly leveraging an always-connected camera to do more than just intelligently make photos look prettier. The app’s actually designed to turn a smartphone’s camera into a document scanner by automating the process of colour-correcting and straightening documents snapped at an angle as well as converting a page’s content to editable text using optical character recognition.
There’s no shortage of these types of apps in either app store, but where iScanner now differentiates itself is a new scanning mode called Count that can intelligently do exactly that based on what it sees in an image.
The process isn’t completely automated through AI, however. Once a photo of a group of objects is snapped you then need to zoom in and highlight an individual object so the app knows what it’s specifically supposed to be counting.
With enough light and contrast in an image, the automated counting is good, but probably a few software updates from being great. It tends to miss objects that aren’t a close visual match to the one you highlighted, but you can also manually add and subtract missed objects or misidentified objects by just tapping the screen. It’s not necessarily faster than counting 24 small bolts by hand, but for larger quantities it can genuinely speed up an otherwise arduous process.
The app gloriously failed the Raymond Babbitt test, however. Even after highlighting a secluded toothpick from a larger pile, iScanner failed to find any visual matches in the pile, ending its career as a potential casino card counter before it even starts. But given how quickly machine learning is evolving, it won’t take long for apps like this to become impressively capable at this skill. I can’t say I’ve ever had the need to count how many toothpicks I’ve spilled, but I’ve definitely got a desk covered in memory cards I’d love to quickly take stock of.