While JPG and PNG metadata extensions (EXIF, IPTC, XMP) allows you to key in various information, the problem here is that there is no pre-defined standards on storing the information of who and where in the photo.
As such, there is no standard way of doing this. While I can key in this info "Crash893:500,300,700,100" in a comment field to denote that Crash893 can be found from pixel coordinates x500,y300 to x700,y100 - there is no guarantee that any program can read it.
AFAIK, FB, Picasa, etc does not store all these info inside the image file itself, relying on external referenced databases (XML, SQL, etc) to store these info.