A new and more powerful engine
Photobook.ai has been raiding our R-and-D labs and quietly released a V2 engine earlier in the year. It radically overhauled our photo analyses engine and removed the dreaded “burst shot” feature. The engine is still just as fast, but way more discerning, especially with similar photos. Detecting similarity in pictures is not rocket science; Computer Vision Ph.D students have been studying this for years. Giving it “human-perception” and to work on mobile photos, however, is – as we found – more challenging than it initially seems. But we cracked it: It’s fast, discerning, human-like and mobile-friendly. This means reframing of shots with shaky hands, HDR or multiple exposures do not fool the algorithm.
Automatic back-cover picture selection
We asked ourselves: How can we use AI to select a back cover image?
So we created PISA: Parting Image Selector Algorithm. We trained a new entire category of images which emotes a “parting shot” or “till the next trip…”
Think long winding roads, sunset over the horizon, feet selfies by the beach… and we are continually adding to this category. So for books with a back cover, we are more likely to pick a photo that will make the end-user go “aww…” and gives closure to the story.
photobook.ai is a spin-off from muvee Technologies, the inventor of automatic video editing and holder of more than 26 patents. Harnessing more than 15 years of deep experience in developing video solutions on mobile coupled with automatic storytelling using computer vision and AI, photobook.ai developed the fastest and smartest photobook creator software on the planet. The photobook.ai SDKs for Automatic Photobook creation (Albumstory) and Advanced Face Recognition and Scoring (iFace) are available for license for both Android and iOS, and also on Windows and Mac OS X.