如何理解caffe的双线性上采样

时间:2016-07-18 07:35:11

标签: caffe bilinear-interpolation

caffe'doc说:

layer {
  name: "upsample", type: "Deconvolution"
  bottom: "{{bottom_name}}" top: "{{top_name}}"
  convolution_param {
    kernel_size: {{2 * factor - factor % 2}} stride: {{factor}}
    num_output: {{C}} group: {{C}}
    pad: {{ceil((factor - 1) / 2.)}}
    weight_filler: { type: "bilinear" } bias_term: false
  }
  param { lr_mult: 0 decay_mult: 0 }
}

我不知道为什么要像这样设置kenrel_size,stride和pad?

1 个答案:

答案 0 :(得分:1)

对于上采样,如果要将大小调整因子设为2,则参数将是kernel_size:4,stride:2,pad:1

相关问题