在视觉皮层中,相同区域的神经元的局部感受野的大小是不同的,从而可以在相

In the visual cortex, the size of the local receptive field of neurons in the same area is different, so that different sizes of spatial information can be obtained in the same processing stage. However, when designing the convolutional network, other attributes of the neuron receptive field have not been considered, such as the adaptive adjustment of the receptive field size. <br>The receptive field size of neurons in the visual cortex is modulated by excitation. In onG, a network with multiple branches, there is a potential mechanism inside which can adjust the Q of the receptive field of the neuron according to the input content in the next convolutional layer. The size, this is because the next convolutional layer merges the features of different branches through linear combination, but this linear combination method is not enough to provide the powerful adjustment ability of the network. SKNet is a non-linear method that fuses features from different nuclei to realize the adjustment of different sizes of the receptive field. It contains three operations: The splt operation generates multiple channels with different nuclear sizes and is related to the different receptive field sizes of neurons; The Fuse operation combines information from multiple channels to obtain a global and understandable representation for weight selection; the Select operation fuses feature maps of different kernel sizes according to the selected weights.

In the visual cortical layer, the size of the local receptor field of neurons in the same region is different, so that spatial information of different sizes can be obtained in the same processing stage. However, other properties of neurons feeling wild were not taken into account when designing the reflow network, such as adaptive adjustments to feel wild size. Visual cortical layer<br>The sensory wild size of the neurons is stimulated to modulate, and in onG, a network with multiple branches, has a potential mechanism to adjust the Q size of the neuron's sensory wild according to the input in the next constgoing layer, because the next constgoing layer fuses the characteristics of different branches by linear combination, but this linear combination is not enough to provide the network's powerful adjustment capability. SKNet is a nonlinear method that combines characteristics from different nucleuts to achieve different size adjustments of the sensory field, which consists of three operations: splt operation produces multiple channels of different nuclear sizes related to the different sensory wild sizes of neurons; Fuse operation combination fuses information from multiple channels to obtain a global and understandable representing for weight selection; Select operations blend feature maps of different core sizes based on the weights selected.

In the visual cortex, the size of the local receptive field of neurons in the same area is different, so different sizes of spatial information can be obtained in the same processing stage. However, when designing convolution network, other properties of receptive field are not considered, such as adaptive adjustment of receptive field size. Visual cortex<br>The size of receptive field of neuron in the next convolution layer is stimulated and modulated. In ong, a network with multiple branches, there is a potential mechanism to adjust the Q size of receptive field of neuron in the next convolution layer according to the input content. This is because the next convolution layer fuses the characteristics of different branches through linear combination, However, this method provides a strong ability to adjust the network. Sknet is a non-linear method to fuse the features from different nuclei to achieve different size adjustment of receptive field. It includes three operations: SPLT operation generates multiple channels with different kernel sizes, which are related to different receptive field sizes of neurons; fuse operation combines the information from multiple channels to obtain a global and understandable representation for weight selection; and fuse operation combines the information from multiple channels to obtain a global and understandable representation for weight selection; The select operation fuses feature maps with different core sizes according to the selected weights.<br>