Real Time Object Identification and depth estimator(both camera-to-object and object-to-object)

The implementation contains three parts:
part 1: object identification, identifies objects within bounding boxes(coco used).
part 2: camera to object distance estimator, identified by frontal face features.
part 3: object to object distance estimator, finds midpoint of each bounding box and calculates the Euclidean distance between two objects.
SSD mobileNet framework of dnn used for part 1 and part 3, for part 2 haar-cascade classifier for frontal face is used.
get the source code and files:
github.com/shi...