For creating personalized Stable Diffusion art, Low Rank Adaption (LoRA) models are all the rage this year. But a new academic offering from China is proposing a system that could bake such elegant functionality directly into Stable Diffusion, without the need to curate data and train models.
You don’t need a powerful GPU to deceive the public – just a mischievous turn of phrase, applied to photos or videos that don’t really support the caption. Now, researchers from Norway are using generative systems such as Stable Diffusion and DALL-E 2 to help root out this kind of ‘cheapfake’ or ‘shallowfake’
New research from the United States and Israel offers a more discrete and less destructive way of editing access to contested material in the Stable Diffusion text-to-image model.
Labeling facial expression data could be helped by the use of large language models such as ChatGPT, and by text/image encoder frameworks such as CLIP. However, these tools are not made for this task, and now new research seeks to adapt them to the challenge.
Open source Facial Emotion Recognition (FER) frameworks are thin on the ground – and what there is, is rather outdated. Now, researchers from USC are proposing a better and lighter FOSS FER architecture, to rival the OpenFace series.
Research from Israel has found an innate quality in Stable Diffusion that may lend itself to producing more temporally consistent video – unlike the usual run of such projects, which ‘bolt on’ complex ancillary frameworks to make text-to-video possible with the world’s favorite generative AI tool
New research from Australia investigates novel ways to teach AI systems how to recognize a far greater range of human facial expressions than just the six basic expressions widely-used in the Facial Action Coding System (FACS). Work of this kind is going to be important very soon, when public awe about neural human recreations in movies and TVs evolves to a higher critical standard.
In the latest contender for stabilizing video output from Stable Diffusion, researchers from China attempt to ‘de-flicker’ generated imagery in the latent space of the system, and achieve some plausible results.
It’s not easy to understand what happens when you make a request to a trained neural network. Whatever the resulting output, if you have no way of understanding what created it, you can’t easily improve or fix the system. Grad-CAM can act as a kind of ‘Barium meal’ that can trace the path of data in output, and highlight it on a heat-map. Now, a new method offers an improvement on the current state-of-the-art, by combining multiple CAM methods with a novel evaluation approach.
Everyone wants temporally coherent Stable Diffusion videos, but the solution has proved elusive. In this later offering, a research team uses the principles of video encoding as a method of getting Stable Diffusion to produce smooth and coherent video.
