Behind the Images – AI Selfies
The images presented in this series are created using a locally developed AI workflow based on Stable Diffusion and ComfyUI. Rather than relying on commercial image-generation platforms, every stage of the process is executed on my own hardware, allowing complete artistic and technical control over the generated images.
The workflow combines custom prompts, personal LoRA models trained on my own photographs, iterative image refinement, and experimental node-based processing in ComfyUI. Each image is the result of numerous iterations, parameter adjustments, and manual decisions rather than a single automated prompt.
The screenshots shown here document parts of this process: the node-based workflow in ComfyUI, the image generation backend, terminal output during model execution, and the custom interface used to control the generation process. Together they reveal the technological infrastructure behind the finished works.
For me, artificial intelligence is not a replacement for photography but an extension of photographic practice. The software becomes a creative instrument that allows me to investigate questions of identity, authorship, memory, and the construction of visual reality.