This study is widely discredited - here is one followup so you can see the difference between ocular (eye) exposure and extraocular (behind the knee):
https://academic.oup.com/jcem/article-pdf/83/9/3369/9231826/jcem3369.pdf
There are indeed visual opsins in the skin, but many of them do not have access to a chromophore, so they don't activate to a neural pathway like those in the eye. There are some potential links with subcutaneous fat in recent experiments, but only in albino mice, and there is no consensus on it yet.