Miscellaneous of the target detection process

---

1. 1*1The role of convolution

1. Dimensionality reduction (dimension reductionality). For example, if a picture with a thickness of 500 * 500 and a depth of 100 is convolved with 1 * 1 on 20 filters, the size of the result is 500 * 500 * 20.

2. Add nonlinearity. After the convolutional layer, it passes through the excitation layer. The 1*1 convolution adds non-linear activation to the learning representation of the previous layer to improve the expressive ability of the network. The 3*3 convolution kernel considers the local Information, 1*1 does not consider local information

references

[1] https://blog.csdn.net/qq_23304241/article/details/80299527