Open parksujin0313 opened 3 years ago
Datasets are just "data". They are not something to be compiled, built, or run on the shell. You will need to use detection models (e.g. YOLO, SSD) and relevant tools (e.g. Keras, TensorFlow, Jupyter) to perform any detection. It seems those scripts in utils/
are based on (or entirely copied from?) those in TensorFlow library, and I suppose they are just some optional helpers that make your validation and conversion easier. You could find how they are used in the tutorial (RoadDamageDatasetTutorial.ipynb
).
@parksujin0313 Hi, Sujin. Sei0o is right. But, if you want more detailed information, email me. I can help you to answer the question.
Hello, I am a Korean student studying artificial intelligence for road defects. I looked it up and there is no overall learning and test function, so I'm leaving a message. I can't see the function that reads and learns the image and sees the result, can you tell me? Or do I have to compile all the functions in Utils?