You need a Restaurant table with a unique primary key REST_ID and other Restaurant attributes (like name and location, name of chef)
You need a Buffet table with the primary key (REST_ID, BUFFET_ID), and other Buffet attributes such as name ( "Salad Bar," "Smorgasbord," or "Sweets" for example.)
You need a Dish table with the primary key (REST_ID, BUFFET_ID, DISH_ID) and other attributes such as name ("Potato Salad," "Buttered Nan")
You need a Photos table with the primary key (REST_ID, BUFFET_ID, DISH_ID, PHOTO_ID) and other attributes, for example a pathname or blob for the photo itself, and a caption.
You need a Reviews table with the primary key (REST_ID, BUFFET_ID, DISH_ID, REVIEW_ID).
Notice that there's a design choice in this data model: it is not possible for the same dish to appear on two different buffets. Nor is it possible for the same buffet to appear in two different restaurants. That is, the data is strictly hierarchical:
One restaurant: zero, one, or more buffets
One buffet: zero, one or more dishes
One dish: zero, one or more photos
One dish: zero, one or more reviews
This makes sense for a review application. If you get food poisoning from the egg salad in Restaurant A, it makes no sense for your negative review of the dish to appear under Restaurant B.
Here are some example rows for each table.
Restaurant (REST_ID, Name, Chef)
1 "McDonalds" Ronald
2 "Julia's Place" Julia
3 "Ritz Carlton Dining Room" Jack
Buffet (REST_ID, BUFFET_ID, Name)
2 1 "Soup Bar" (this is the soup bar at Julia's place)
2 2 "Salad Bar" (the salad bar at Julia's place)
3 1 "Cold Breakfast" (the cold breakfast bar at the hotel dining room)
3 2 "Sweets" ( the dessert bar at the hotel dining room)
Dish: (REST_ID, BUFFET_ID, DISH_ID, name and category)
2 1 1 "Minestrone" "Soup"
2 1 2 "French Onion" "Soup"
2 1 3 "Vegetarian Barley" "Soup"
2 1 4 "Saltines" "Crackers"
3 1 1 "Melon" "Fruit"
3 1 2 "Strawberry" "Fruit"
3 1 3 "Blueberry Muffin" "Bread"
3 1 4 "Multigrain Bread" "Bread"
3 1 5 "Corn Flakes" "Cereal"
3 1 6 "Orange Juice" "Drink"
3 1 7 "Milk" "Drink"
etc.