Deep affordance learning for single- and multiple-instance object detection

Affordance learning in general, is to identify the purpose, use, and ways to interact with an object, based on information gained from observing the object. Most of the existing affordance learning approaches assume the object target has been cropped individually from images. However, the object could not be easily separated from others due to occlusion or noise. Actually, two or more neighboring objects belong to the same class could be detected as one object target. A fault affordance may thus result. In this paper, we propose an extension of the existing object detection by adding a classifier which can recognize the object to two cases: single instance and multiple instances. By doing so, the concept of affordance learning was introduced to utilize the visual information from detected instances and to understand other properties of the object such as if it is singular or multiple, upright or tilted, rigid or deformable, an even movable or static. The algorithm is implemented in Python to provide a one-stop solution for dataset building and management. Experimental results have been enclosed to show the effectiveness and accuracies of the proposed approach.