Viper Vision 1: Fundamental physics of biological vision systems

Biological vision systems operate at the extremes of temporal response, spatial acuity, and (potentially single quanta) sensitivity, yet fundamental understanding of these extraordinary capabilities is lacking. Here, we develop a generalized physical model of biological vision systems based on fundamental aspects of noisy light inputs. As a key application of our model, we study the infrared vision in pit vipers. We show that a vision system optimized to assess thermal fluctuations at each pixel in space is remarkably efficient at finding boundaries between regions of differing temperatures. By optimizing this edge-finding capability, we show that adaptive, neuromorphic systems are exceptionally efficient and robust, especially in terms of hardware simplicity and computational efficiency. We compare our proposed vision system to state-of-the-art infrared materials and devices and show that water—a biologically relevant infrared absorber—is in fact ideal for natural infrared vision systems. Beyond explaining several long-standing questions in the pit viper vision community, our model has the potential to expose new understanding in vision systems ranging from dragonflies and honey bees to mantis shrimp and humans.