A method for image proces in neural network

The invention provides an image processing method in a neural network. The method comprises the following steps: a first classification network performs feature extraction on the received target image; a second classification network performs feature extraction on the received target image. The first positioning network transmits the obtained first target pixel number and the first target positionto the first classification network. The first classification network determines a pooling result of the selected first target pooling layer according to the number of the first target pixels. A second classification network performs feature extraction on each of the first plurality of feature maps; The second classification network determines a pooling result of the selected second target pooling layer to obtain a second prediction tag of the second scale layer; Identify the target category label and determine the final target location. In the embodiment of the invention, a plurality of a priori rectangular frames a