Dynamic Cloud Offloading for 2D-to-3D Conversion

In this paper, a dynamic offloading model together with a cloud-friendly depth estimation algorithm is proposed to minimize the energy consumption of mobile devices by exploiting cloud computational resources for 2D-to-3D conversion. The cloud-friendly depth estimation algorithm partitions an input image into several parts, classifies each part to a specific type, and applies a specific conversion algorithm to each type to generate depth maps, which facilitates allocating the partitions between the mobile device and the cloud dynamically. Then, a dynamic offloading model is proposed for mobile energy minimization by allocating the partitions to be processed dynamically between the cloud and the mobile. The complexity of depth estimation, the processing capability of the cloud, and the power consumption of the mobile are considered jointly into the model to provide an optimized solution. Several simulations based on parameters of real mobile devices demonstrate that our method can save an average of almost 21.17% of total energy on different mobile devices and an average of 17.09% of total energy under different transmitting rates than the existing algorithms for 2D-to-3D conversion.