Image data blur processing through object detection

Abstract

Object detection is useful for understanding the content of an image by describing both the content of the image and the location of that object. The goal of object detection is to recognize instances of a set of predefined object classes and use bounding boxes to describe the position of each detected object in the image. There are two different approaches for this task. Fixed number of predictions on grid (one stage) or leverage a proposal network to find objects and then use a second network to fine-tune these proposals and output a final prediction (two stage).

In the case of blur processing, processing by hand is common. In order to reduce such inconvenience, it was conceived that it would be efficient if the detection process could be performed together with the blur processing at once. If such processing is possible, it may be applicable to various image data.