Efficient Image Sensor Subsampling for DNN-Based Image Classification

Today's mobile devices are equipped with cameras capable of taking very high-resolution pictures. For computer vision tasks which require relatively low resolution, such as image classification, subsampling is desired to reduce the unnecessary power consumption of the image sensor. In this paper, we study the relationship between subsampling and the performance degradation of image classifiers that are based on deep neural networks (DNNs). We empirically show that subsampling with the same step size leads to very similar accuracy changes for different classifiers. In particular, we could achieve over 15x energy savings just by subsampling while suffering almost no accuracy lost. For even better energy accuracy trade-offs, we propose AdaSkip, where the row sampling resolution is adaptively changed based on the image gradient. We implement AdaSkip on an FPGA and report its energy consumption.