About these AI images ...
I have been fascinated by the AI images various ppl are posting to their blogs. The techniques and tips posted by cybertied led me to give Bing a try. I found it easy and fast to create the kind of images that interest me. I also found it frustrating.
Having no experience with this kind of thing, I was able to successfully create some 50+ images in less than two hours spread over two days. I was also blocked many times while trying to create an image that seemed innocuous and similar to others I had created. The final straw happened when I was suspended for an hour trying to create what I thought was a relatively simple image. It is not clear what Bing thinks violates their TOS. IMO, this technology is too important to be left to the likes of Microsoft. I have since deleted my Bing / Microsoft account. My curiosity has been satisfied. I don't know that I will try doing more of this AI image creation anytime soon.
Below are some of the images I created. The starting point for many of them was "marine prisoner wearing desert mar pat uniform locked inside jail cell." Most times Bing would give variations of what it thinks were close approximations to what I asked for. Other times, Bing would mis-apply something I specified … like substituting brown "navy" uniforms when I had specified "dress white sailor" uniforms. In general, I have been quite impressed by the quality of the rope work other ppl have obtained in their images. That said, when cuffs / irons are asked for, the results are uniformly disappointing; the steel restraints shown are not realistic and leave a lot to be desired. Certainly there are more than enough civilian / military images on the net showing restraints in use that are identifiable as to their maker, construction, proper use, etc. The ppl doing this AI work need to up their game when it comes to images of cuffs / irons. IMO.