A camera capable of automatically detecting a main object in a sensed image. A distance distribution generation unit discriminates a plurality of areas in the sensed image on the basis of a predetermined condition, and a main object area determination unit automatically determines the main object area out of the discriminated areas. If the determined main object area is different from the area where an operator intends to focus on and the operator wishes to change the main object area, a new main object area is selected in accordance with an instruction inputted from a main object area changing unit. Then, the camera focuses on the selected main object area.