Institutional policies must be followed, if mention the permission process for using photos of students; and ethical principles of informed written consent from students, where no element of coersion is present; only, then one can use the photos of students and some other persons.
I guess it depends on the publisher. Some journals do not have specific ruling on it as long as the students were informed and consent was obtained. Plus, you're not showing them doing something harmful.
While it may be true that we may not be intentionally showing a research participant doing something harmful (although I have way too many times seen published pictures of learners in a science education lesson who are NOT wearing adequate safety equipment such as googles, etc, and thus are in danger), one never knows if any particular student could be inadvertently placed in some type of danger by revealing her/him in a published picture that is available for facial identification techniques. Odd circumstances exist, as we all know, when we deal with real people in classroom settings. Therefore, I am in complete support of the key principle of "Do No Harm," so I always request written permission from all individuals before including such pictures in my published research. Agree, too, with Ian Kennedy that fuzzed out facial features is not an agreeable option!