While framing research has centered mostly on the evaluations of media texts, visual news discourse has remained relatively unexamined. This study surveys the visual framing techniques and methods employed in previous studies and proposes a four-tiered model of identifying and analyzing visual frames: (1) visuals as denotative systems, (2) visuals as stylistic-semiotic systems, (3) visuals as connotative systems and (4) visuals as ideological representations. These four tiers are defined and the process of identifying frames at each level is explicated. The proposed system can be applied to analyzing any type of visual media content or audiences' perception of that content.