I am running the NVIDIA-Optical-Character-Detection-and-Recognition-Solution locally on my Jetson. I want to block the pipeline of OCDNet and only infer with OCRNet. I commented out all the code related to OCDNet. By the result is coming out as empty inference.
I solved it by myself. The OCDNet output it’s detection as polygon points. Then the output of OCDNet is inserted into OCRNet as input. I created a polygon which represents the whole image and inserted it into OCRNet and blocked the OCDNet polygons.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.