Revising deep learning methods in parking lot occupancy detection

Parking guidance systems have recently become a popular trend as a part of the smart cities' paradigm of development. The crucial part of such systems is the algorithm allowing drivers to search for available parking lots across regions of interest. The classic approach to this task is based on the application of neural network classifiers to camera records. However, existing systems demonstrate a lack of generalization ability and appropriate testing regarding specific visual conditions. In this study, we extensively evaluate …

algorithm application become cities cs.cv cs.lg deep learning detection development guidance paradigm parking part popular search smart smart cities systems the algorithm trend