I. Introduction
The last few decades have seen an explosion in the use of cameras and other imaging devices. It is estimated that in the year 2017 alone, more than a billion camera modules were sold, most of them integrated into mobile systems such as phones, tablets, and other devices. The miniaturization and rapidly decreasing cost of camera modules have been the primary drivers of this scaling. With the rise in use and integration of cameras, their role in our lives has also changed significantly. Most cameras no longer take photographs, but instead are used as sensors to provide inferential inputs for a diverse range of applications, from biometrics (face recognition, authentication) to surveillance and security. While modern camera modules can be thin ( mm) and inexpensive ()
Estimates are based on breakdown analyses of the camera modules in current mobile phones [1].
, many emerging applications, such as the Internet of Things (IoT), surveillance, disaster recovery using drones, etc., impose stringent constraints on size, weight, and cost that cannot be met with today's lens-based imaging systems. Consider the integration of cameras into household electronics and electric appliances, such as thermostats, coffee machines, toasters, refrigerators, etc. Given the cost-constraints in these devices, integration is only feasible if we can realize an order of magnitude reduction in the cost of camera modules. Consider also the integration of cameras into drones for distributed sensing in applications such as disaster recovery or in very thin objects such as credit cards. Given the weight and volume constraints in these devices, integration is only possible if we can realize an order of magnitude reduction in the weight and size of camera modules (see Fig. 1).