Spectalix developed a unique video creator for businesses around the world as the ultimate user engagement solution for their consumer apps in a way not seen before. It is being offered as a customizable white label app or as an SDK to be integrated into their mobile app. It allows mobile phone users to separate people from their background, using either live video capture or pre-recorded clips in the phone gallery. The user can then place the segmented person in different branded video clips, which are constantly being uploaded onto the brand’s app, and interact with its talents in a very engaging and fun way.
The app offers a special editor which allows the user to resize and reposition the video of the cut-out person, in order to place it perfectly in the context of the background template. Business now have the tools to dramatically increase user engagement and effectively monetize their brand and talents. That said, the large exposure on social media apps will be kept intact, as users will share their disruptive videos on social media and expose the brand and its app to many more eyeballs.
Our core technology is based on a super-compact segmentation neural-network (NN), which recognizes multiple people and other moving objects from every angle or position, segments these objects in real-time and then draws them on a new background, be it an image or a video clip.
The network is optimized to run on the limited computing capabilities of mid-range mobile chipsets and devices, thanks to a patent pending method which combines the NN with highly parallelized computer vision algorithms that increase the overall performance.
Spectalix segmentation is done on real-time captured video, in full HD and 30fps. It is 100% software based, cross platform and requires only a single 2D camera (no need for depth data and h/w sensors). Spectalix is also using a high quality, home-grown dataset, specifically tailored for real-time object segmentation in video on mobile phones.
Our core technology is based on a super-compact segmentation neural-network (NN), which recognizes multiple people and other moving objects from every angle or position, segments these objects in real-time and then draws them on a new background, be it an image or a video clip.
The network is optimized to run on the limited computing capabilities of mid-range mobile chipsets and devices, thanks to a patent pending method which combines the NN with highly parallelized computer vision algorithms that increase the overall performance.
Spectalix segmentation is done on real-time captured video, in full HD and 30fps. It is 100% software based, cross platform and requires only a single 2D camera (no need for depth data and h/w sensors). Spectalix is also using a high quality, home-grown dataset, specifically tailored for real-time object segmentation in video on mobile phones.
Creative Director at Academy award winning Film Production company in addition to overseeing the TV license business. Visionary entrepreneur and consultant in the Israeli high-tech industry.
Serial entrepreneur (founder of Vodience). Acted as CTO of OTT TV services such as Cellcom TV and BlockbusterTV in Israel. Author of the Novel “Twice life, please”
Deep Learning algorithm expert with special focus in embedded devices. Ex-developer at Advanced Imaging Technology Group of Microsoft. B,Sc in Computer Engineering at Technion, M.Sc in Computer science at Haifa University.
Various roles of RnD, leading and growing development teams from initial phase to 30 people. Founded Vodience. Among the core team that founded Press-Sense (Raised $18M USD). B.Sc from the Technion.
Subscribe to our newsletter