Using ControlNet and InstantID, faces can be edited in multiple ways with GenAI. Below is an example using an SDXL model and various off-the-mill prompts, leveraging the „SDXL Styles“ extension for Automatic1111.
We are all experimenting with GenAI and its capabilities on a daily basis. The alignment of real-world use cases with AI capabilities will play a central role in 2024 as we move forward. It’s something every company is currently dealing with. Where I was asked to hold talks about the sheer potential of AI in the past year, people are now demanding closer guidance, workshops, promptathons (right Sindy?) and mentorships when dealing with the challenges ahead.
But let me also be clear about leveraging (Gen)AI: just because a technology suddenly exists doesn’t make it better or an automatic priority to be integrated into legacy workflows and problem-solving. Instead, we should focus on efficiency, smooth workflows and easy implementation while looking at what tool or process is best fit to do the job. This should take center stage as we progress.
AI image production surely has come a long way since the „early“ days in 2022.