Risky Region Localization With Point Supervision

Kazuki Kozuka, Juan Carlos Niebles; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 246-253

We propose a method for detecting regions with potential risk from images. We focus on images acquired by a front camera mounted on a car with the goal of localizing image regions where pedestrians are likely to enter the scene suddenly. In this case, we define the risk value at every pixel as the likelihood that a pedestrian will occupy those pixels shortly. This task is very challenging because the risk areas are not easily characterized by appearances of single objects, and therefore these regions exhibit large visual variations. Additionally, the boundaries of the risk regions in the image are not easily defined by human annotators, as they do not tend to correspond to object boundaries. This causes the annotation process to be ambiguous and costly. Instead of relying on ambiguous annotations of the boundaries of risk regions, we propose a weakly supervised method for risk region localization and risk value estimation that only requires 1 point supervision at training time.