The new search feature can now produce results based on subjective descriptions of photos, as opposed to just tags. Image: INQUIRER.net stock photo
Facebook is getting a backend update regarding its search feature, thanks to its computer vision platform. Users will now be able to search for photos based on the contents of the image.
The Verge reports that keywords describing the contents of a photo will now highlight posts that match the description, no photo tags necessary.
These search results will place higher in the grid with other relevant photos listed after them. The functionality is similar to the image-recognition feature on Google Photos or iOS 10. For example, a keyword search of “pictures of pizza” will show photos from friends and public results that feature pizza in them, be it ad, group photos, or memes.
Facebook also appears to tie search results from other context clues, such as captions and comments.
Image: INQUIRER.net
The update is available to mobile and web users in the U.S. but some local users in the Philippines may also be able to use the new search feature. This aligns with Facebook’s efforts to integrate computer vision technology into consumer-friendly products. TechCrunch notes that eventually, the feature will also be applied to video searches.
The funny thing about this feature is that it even recognizes slang keywords—say, ‘waifu,’—but will produce varied results when searching for ‘cute pictures of me.’ Alfred Bayle