Real-time detection of sepsis on a video data is a new aboard technique that aids the septic patient and decreases the high mortality rate. The progressive impairment of the micro-circulation associated with increased systemic inflammatory response in sepsis has been considered the origin of the multiple organ dysfunction syndrome that often leads to death. However, despite the recognized importance of the micro-circulatory dysfunction, analysis methods able to correlate the severity of sepsis with the degree of impairment of micro-hemodynamic captured by portable microscope Side-stream Dark Field Imaging (SDF) are rarely used. Hence, the classification of the severity of sepsis by analyzing the micro-circulatory dysfunction would be of great assistance in diagnosing severity and therapeutic management. In this context, the aim of this work is to propose a new computational methodology based on image processing to obtain graph metrics for determining the degree of micro-vascular and tissue commitment due to sepsis.