Our approach towards managing the occlusion problem is to introduce camera pixels consisting of a two-device tandem structure. The first layer carries out object detection, in the following referred to as object detection layer ODL, utilizing any standard image capturing pixels such as silicon photodiodes. The second layer, which is the focus of this article, handles the correct identification of occluded objects and is hence termed occlusion handling layer OHL, as schematically depicted in Figure 1a. Both sensing layers operate independently, in this way enabling fast parallel computing. We first investigate the simplest scenario involving a stationary object being masked by another body passing by as the foreground (Movie 1 simulates a simple example where a stationary bright square is being blocked by a crossing red square). In this situation, our AI accelerating sensor carries out object detection using the following algorithm (Figure 1b):