In OpenVINO R3, there is a feature to mark the input image as resizable which allows the inference engine to accept image of any size.
InputInfo::Ptr input_info = network.getInputsInfo().begin()->second; input_info->getPreProcess().setResizeAlgorithm(RESIZE_BILINEAR);
I would like to know if this option is made available in python API also. If so, could you please let me know how to do it as I could not find it anywhere in the documentation.
In OpenVINO R3 release pre-processing functionality available only in C++ Inference Engine API. We are working on aligning features between C++ and Python API and in future releases pre-processing functionality will be available in Python API also.