The Ground Truth #9

This year, we gave thanks for new technology, and lots of it. We’ve brought you all the latest news from the industry so scroll down to read more…

From the Industry

It’s been a busy time for conferences. In late October, ECCV 2022 was held in Tel Aviv and the conference was jam-packed with discoveries in the computer vision field. You can read the highlights from ECCV 2022 here and here. We covered DepthGAN, Make-a-Scene, editing the lighting of an indoor scene from a single image, Text2Live, Grasp’D and regenerating a scene from a novel perspective or lighting condition.??

?Happening this month is NeurIPS 2022. We reviewed 5 breakthroughs in synthetic data being presented at the conference ranging from “Generalized one-shot domain adaptation of generative adversarial networks” to “D2NeRF: Self-Supervised Decoupling of Dynamic and Static Objects from a Monocular Video” and more. You can read about all 5 papers (and even some bonus ones!) here.?

From Academia

Aging is a fascinating challenge for the computer vision field and for face recognition. As part of our ongoing series on facial recognition, we’ve reviewed the paper on:??

Custom Structure Preservation in Face Aging

Face age transformation (editing) involves reflecting age factors on a given face to create its future or past face images. A transformed face image must retain the identity of the original photo subject and all other details related to the face image.?

?Over the last few years, this problem has attracted increasing attention, both in industry and academia, due to its wide range of applications. Despite extensive research on face aging, the lack of training samples for a given person over a long period of time still makes face aging a very difficult task.

?The authors developed a model capable of performing a realistic face image transformation successfully reflecting face shape changes happening when a person ages. The researchers observed that different individuals age differently and in a non-deterministic fashion. Some people's appearances change significantly over time, however, others can easily be recognized from old photographs. To reflect this phenomenon, the authors developed a masking mechanism which allows for degree of structure preservation to be adjusted during inference.?

?Read the paper review here.

From the Team

Introducing Our 50K Identities

We just released our new 50K face identities! Now you can choose from endless permutations using our new photo-realistic synthetic identities to further improve your CV models. These new identities will enable our customers to bring better CV models to production faster. Each identity is free from privacy concerns, its uniqueness is algorithmically proven and fully controllable on a granular level by using our SDK. Learn more here or get started now! Learn more here or get started now!

Fresh New Content

New Podcast - SLAM and the Evolution of Spatial AI

Host and Datagen CTO and co-founder, Gil Elbaz welcomes Andrew J. Davison, the father of SLAM. Andrew and Gil dive right into how SLAM has evolved and how it started. They speak about Spatial AI and what it means along with a discussion about global belief propagation. Of course, they talk about robotics, how it's impacted by new technologies like NeRF and what is the current state-of-the-art. Listen now.?