Adjust the volume
Have a seat.\ So.\ Half of these works were created by mixing the faces that were generated by the thispersondoesntexist website and the generic person blanks that I created in Adobe Fuse. Thispersondoesntexist at that time was one of the first available “technologies” of this kind.
Afterwards in Adobe Photoshop I placed the “face of the neural network” and put a model under it. I was playing around with the light and shadows, making then the post-processing, using various graphic elements found on Google-images.
\ In general, it turned out that I was attaching a certain context to the lifeless faces of the neural network, reviving them, and as a result, a new meaning was being made. The scope of the context has always been set intuitively and completely random.
After Artbreeder, everything acquired a completely different course. I began to pass my own face through the neural network, thereby generating a greater number of new images. But again, I did not dwell on the final images that the AI gave out to me.
\ Everything was going the same way, but using 3D this time was excluded. I was using post-production - mixing faces with various graphic elements (textures, backgrounds, objects). Very often, I took generic “samples” (eyes, ears, mouths, noses) from my old previous works.