How to read this: the model compares your image with the exact candidate names you supply. Add plausible alternatives for a useful comparison. Results are exploratory, not expert identification.