Open hainguyenct opened 8 years ago
I see convolution in the code below to classify faces: (image 32x32). He used nstates={16, 32}, filtersize={5,7}, poolsize=4 for image 32x32
These parameters are for image 32x32. I am wondering what number we could use if the size of image are 128x128, 256x256,...? Please help me. Thank you
I see convolution in the code below to classify faces: (image 32x32). He used nstates={16, 32}, filtersize={5,7}, poolsize=4 for image 32x32![screen shot 2016-05-23 at 11 46 26](https://cloud.githubusercontent.com/assets/17986080/15467012/84d36f30-20dc-11e6-9c5e-789499d6c1ea.jpg)
These parameters are for image 32x32. I am wondering what number we could use if the size of image are 128x128, 256x256,...? Please help me. Thank you