PytorchSSD
PytorchSSD copied to clipboard
How is the detection effect on small targets?
How is the detection effect on small targets?
I trained this model (SSD_vgg) in the dataset of "GDXray" which composes of many small objects. But the mAP is not good. I suggest the input size of the image and the size of the prior box need to be adjusted.
You should use large images to handle the small targets.