I don't know the original source material enough to tell if this is good or not, but from the looks of the results I think "high tops" may be a better tag to use in training over much the more generic "sneaker". It's a tag Illustrious (and Anima) already recognize.

Also if you don't know you can use tags like "photo (medium), lower body, footwear focus" or "photo (medium), unworn shoes, footwear focus, still life" when training off product images. The only real problem it causes for Illustrious is that if that's all /most of your data generations may need to specify facial attributes/actions (something the overwhelming majority of gens do anyways), or use "lower body" as a negative to overcome the default camera. Edit: If you're doing footwear LoRAs, "shoe soles" is also a recognized tag and useful if your subject has unique tread.