2026 ASEE Annual Conference & Exposition

Object detection systems deployed in campus surveillance environments face documented failure modes. Crowd density causes occlusion. Poor camera quality degrades features. Variable lighting obscures objects. Motion blur distorts boundaries. Temporal anomalies shift baseline expectations. The computer vision literature identifies each factor independently, yet no unified framework exists to quantify their combined effect on detection difficulty. This paper synthesizes findings from peer-reviewed studies to propose the Campus Complexity Index (CCI), a five-component framework combining established failure modes into a single computable metric. We describe the literature selection methodology, the coding process for extracting failure modes, the mathematical formulation of each component, and the rationale for their combination. To demonstrate feasibility, we validate on two public datasets: UCF CC 50 (50 crowd images) and the Monash Guns Dataset (7,780 weapon detection images), showing that CCI produces meaningful variation between scenario types. Full empirical validation comparing detection performance with CCI-adaptive versus static thresholds remains future work.