One sample image belongs to my dataset. I was planing to use YOLO for detect weeds on corn farm. But, as you see, it is quite complex. Other images even more complex. Difficult to precisely annotate. So, I had decided to create a classification tflite model and use sliding window aproach. Now, it works at "not bad" level, but its too slow. Do you have any advice or guide for my situation? Thank you.