A library for performing object detection utilizing the Single Shot MultiBox Detector model created by google.
Input for aperature takes a type implementing the DetectionImage trait and performs object detection on that image. The list of objects that can be detected via the label map for the model can be found here
Output is returned as a hash map of the detection tensors containing the detection boxes, detection classes detection scores, and the number of detections. Each element of the returned HashMap can be accessed by the name of the tensor in all lower case separated by an underscore. (e.x. detection_scores)
Image module representing the input image to be inferenced.
ObjectDetection module that handles the setup and inferencing of the image data.