EgoNet_RGB_DHG.prototxt

#prototxt file
name: "EgoNet"

#data
input: "data"
input_dim: 1
input_dim: 3
input_dim: 481
input_dim: 641

#data
input: "dhg_data"
input_dim: 1
input_dim: 3
input_dim: 481
input_dim: 641

input: "spatial_data"
input_dim: 1
input_dim: 3
input_dim: 481
input_dim: 641

layer {
  name: "slicer_data"
  type: "Slice"
  bottom: "spatial_data"
  top: "Y"
  top: "X"
  top: "junk"
  slice_param {
    slice_point: 1
    slice_point: 2
  }
}


### NETWORK ###

layer {
  bottom: "data"
  top: "conv1_1"
  name: "conv1_1"
  type: "Convolution"
    convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }

  }
}

layer {
  bottom: "conv1_1"
  top: "conv1_1"
  name: "relu1_1"
  type: "ReLU"
}

##

layer {
  bottom: "dhg_data"
  top: "dhg_conv1_1"
  name: "dhg_conv1_1"
  type: "Convolution"
  convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }

  }
}

layer {
  bottom: "dhg_conv1_1"
  top: "dhg_conv1_1"
  name: "dhg_relu1_1"
  type: "ReLU"
}

##

layer {
  bottom: "conv1_1"
  top: "conv1_2"
  name: "conv1_2"
  type: "Convolution"
  convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv1_2"
  top: "conv1_2"
  name: "relu1_2"
  type: "ReLU"
}
###

layer {
  bottom: "dhg_conv1_1"
  top: "dhg_conv1_2"
  name: "dhg_conv1_2"
  type: "Convolution"
  convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv1_2"
  top: "dhg_conv1_2"
  name: "dhg_relu1_2"
  type: "ReLU"
}
###

layer {
  bottom: "conv1_2"
  top: "pool1"
  name: "pool1"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}

layer {
  bottom: "pool1"
  top: "conv2_1"
  name: "conv2_1"
  type: "Convolution"
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv2_1"
  top: "conv2_1"
  name: "relu2_1"
  type: "ReLU"
}

###

layer {
  bottom: "dhg_conv1_2"
  top: "dhg_pool1"
  name: "dhg_pool1"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}

layer {
  bottom: "dhg_pool1"
  top: "dhg_conv2_1"
  name: "dhg_conv2_1"
  type: "Convolution"
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv2_1"
  top: "dhg_conv2_1"
  name: "dhg_relu2_1"
  type: "ReLU"
}

####

layer {
  bottom: "conv2_1"
  top: "conv2_2"
  name: "conv2_2"
  type: "Convolution"
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv2_2"
  top: "conv2_2"
  name: "relu2_2"
  type: "ReLU"
}
layer {
  bottom: "conv2_2"
  top: "pool2"
  name: "pool2"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}

####

layer {
  bottom: "dhg_conv2_1"
  top: "dhg_conv2_2"
  name: "dhg_conv2_2"
  type: "Convolution"
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv2_2"
  top: "dhg_conv2_2"
  name: "dhg_relu2_2"
  type: "ReLU"
}
layer {
  bottom: "dhg_conv2_2"
  top: "dhg_pool2"
  name: "dhg_pool2"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}
####

layer {
  bottom: "pool2"
  top: "conv3_1"
  name: "conv3_1"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv3_1"
  top: "conv3_1"
  name: "relu3_1"
  type: "ReLU"
}

###

layer {
  bottom: "dhg_pool2"
  top: "dhg_conv3_1"
  name: "dhg_conv3_1"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv3_1"
  top: "dhg_conv3_1"
  name: "dhg_relu3_1"
  type: "ReLU"
}

####

layer {
  bottom: "conv3_1"
  top: "conv3_2"
  name: "conv3_2"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv3_2"
  top: "conv3_2"
  name: "relu3_2"
  type: "ReLU"
}

###

layer {
  bottom: "dhg_conv3_1"
  top: "dhg_conv3_2"
  name: "dhg_conv3_2"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv3_2"
  top: "dhg_conv3_2"
  name: "dhg_relu3_2"
  type: "ReLU"
}
####

layer {
  bottom: "conv3_2"
  top: "conv3_3"
  name: "conv3_3"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv3_3"
  top: "conv3_3"
  name: "relu3_3"
  type: "ReLU"
}

###
layer {
  bottom: "dhg_conv3_2"
  top: "dhg_conv3_3"
  name: "dhg_conv3_3"
  type: "Convolution"
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv3_3"
  top: "dhg_conv3_3"
  name: "dhg_relu3_3"
  type: "ReLU"
}
####

layer {
  bottom: "conv3_3"
  top: "pool3"
  name: "pool3"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}

###

layer {
  bottom: "dhg_conv3_3"
  top: "dhg_pool3"
  name: "dhg_pool3"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
    pad: 1
  }
}

###

layer {
  bottom: "pool3"
  top: "conv4_1"
  name: "conv4_1"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}

layer {
  bottom: "conv4_1"
  top: "conv4_1"
  name: "relu4_1"
  type: "ReLU"
}

###

layer {
  bottom: "dhg_pool3"
  top: "dhg_conv4_1"
  name: "dhg_conv4_1"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}

layer {
  bottom: "dhg_conv4_1"
  top: "dhg_conv4_1"
  name: "dhg_relu4_1"
  type: "ReLU"
}

###

layer {
  bottom: "conv4_1"
  top: "conv4_2"
  name: "conv4_2"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv4_2"
  top: "conv4_2"
  name: "relu4_2"
  type: "ReLU"
}
###

layer {
  bottom: "dhg_conv4_1"
  top: "dhg_conv4_2"
  name: "dhg_conv4_2"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv4_2"
  top: "dhg_conv4_2"
  name: "dhg_relu4_2"
  type: "ReLU"
}

###

layer {
  bottom: "conv4_2"
  top: "conv4_3"
  name: "conv4_3"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "conv4_3"
  top: "conv4_3"
  name: "relu4_3"
  type: "ReLU"
}
###

layer {
  bottom: "dhg_conv4_2"
  top: "dhg_conv4_3"
  name: "dhg_conv4_3"
  type: "Convolution"
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv4_3"
  top: "dhg_conv4_3"
  name: "dhg_relu4_3"
  type: "ReLU"
}
###

layer {
  bottom: "conv4_3"
  top: "pool4"
  name: "pool4"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    #pad: 1
    #stride: 2
    stride: 1
  }
}

###

layer {
  bottom: "dhg_conv4_3"
  top: "dhg_pool4"
  name: "dhg_pool4"
  type: "Pooling"
  pooling_param {
    pool: MAX
    kernel_size: 2
    #pad: 1
    #stride: 2
    stride: 1
  }
}

###


layer {
  bottom: "pool4"
  top: "conv5_1"
  name: "conv5_1"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "conv5_1"
  top: "conv5_1"
  name: "relu5_1"
  type: "ReLU"
}

###

layer {
  bottom: "dhg_pool4"
  top: "dhg_conv5_1"
  name: "dhg_conv5_1"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv5_1"
  top: "dhg_conv5_1"
  name: "dhg_relu5_1"
  type: "ReLU"
}
###


layer {
  bottom: "conv5_1"
  top: "conv5_2"
  name: "conv5_2"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "conv5_2"
  top: "conv5_2"
  name: "relu5_2"
  type: "ReLU"
}
###
layer {
  bottom: "dhg_conv5_1"
  top: "dhg_conv5_2"
  name: "dhg_conv5_2"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv5_2"
  top: "dhg_conv5_2"
  name: "dhg_relu5_2"
  type: "ReLU"
}
###


layer {
  bottom: "conv5_2"
  top: "conv5_3"
  name: "conv5_3"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "conv5_3"
  top: "conv5_3"
  name: "relu5_3"
  type: "ReLU"
}
####

layer {
  bottom: "dhg_conv5_2"
  top: "dhg_conv5_3"
  name: "dhg_conv5_3"
  type: "Convolution"
  convolution_param {
    num_output: 512
    #pad: 1
    pad: 2
    dilation: 2
    kernel_size: 3
  }
}
layer {
  bottom: "dhg_conv5_3"
  top: "dhg_conv5_3"
  name: "dhg_relu5_3"
  type: "ReLU"
}
###

layer {
  bottom: "conv5_3"
  top: "pool5"
  name: "pool5"
  type: "Pooling"
  pooling_param {
    pool: MAX
    #kernel_size: 2
    #stride: 2
    kernel_size: 3
    stride: 1
    pad: 1
  }
}

###

layer {
  bottom: "dhg_conv5_3"
  top: "dhg_pool5"
  name: "dhg_pool5"
  type: "Pooling"
  pooling_param {
    pool: MAX
    #kernel_size: 2
    #stride: 2
    kernel_size: 3
    stride: 1
    pad: 1
  }
}

###


layer {
  bottom: "pool5"
  top: "fc6"
  name: "fc6"
  type: "Convolution"
  convolution_param {
    num_output: 4096
    pad: 6
    dilation: 4
    kernel_size: 4
  }
}
layer {
  bottom: "fc6"
  top: "fc6"
  name: "relu6"
  type: "ReLU"
}
layer {
  bottom: "fc6"
  top: "fc6"
  name: "drop6"
  type: "Dropout"
  dropout_param {
    dropout_ratio: 0.5
  }
}

####

layer {
  bottom: "dhg_pool5"
  top: "dhg_fc6"
  name: "dhg_fc6"
  type: "Convolution"
  convolution_param {
    num_output: 4096
    pad: 6
    dilation: 4
    kernel_size: 4
  }
}
layer {
  bottom: "dhg_fc6"
  top: "dhg_fc6"
  name: "dhg_relu6"
  type: "ReLU"
}
layer {
  bottom: "dhg_fc6"
  top: "dhg_fc6"
  name: "dhg_drop6"
  type: "Dropout"
  dropout_param {
    dropout_ratio: 0.5
  }
}

####

layer {
  bottom: "fc6"
  top: "fc7"
  name: "fc7"
  type: "Convolution"
  convolution_param {
    num_output: 4096
    kernel_size: 1
  }
}
layer {
  bottom: "fc7"
  top: "fc7"
  name: "relu7"
  type: "ReLU"
}
layer {
  bottom: "fc7"
  top: "fc7"
  name: "drop7"
  type: "Dropout"
  dropout_param {
    dropout_ratio: 0.5
  }
}
###

layer {
  bottom: "dhg_fc6"
  top: "dhg_fc7"
  name: "dhg_fc7"
  type: "Convolution"
  convolution_param {
    num_output: 4096
    kernel_size: 1
  }
}
layer {
  bottom: "dhg_fc7"
  top: "dhg_fc7"
  name: "dhg_relu7"
  type: "ReLU"
}
layer {
  bottom: "dhg_fc7"
  top: "dhg_fc7"
  name: "dhg_drop7"
  type: "Dropout"
  dropout_param {
    dropout_ratio: 0.5
  }
}

###


layer {
  bottom: "X"
  top: "X_shrink"
  name: "X_shrink"
  type: "Interp"
  interp_param {
    shrink_factor: 8
    pad_beg: -1
    pad_end: 0
  }
}

layer {
  bottom: "Y"
  top: "Y_shrink"
  name: "Y_shrink"
  type: "Interp"
  interp_param {
    shrink_factor: 8
    pad_beg: -1
    pad_end: 0
  }
}

layer {
  name: "fc8_concat"
  bottom: "fc7"
  bottom: "dhg_fc7"
  bottom: "X_shrink"
  bottom: "Y_shrink"
  top: "fc8_concat"
  type: "Concat"
}


layer {
  bottom: "fc8_concat"
  top: "fc9"
  name: "fc9"
  type: "Convolution"
  convolution_param {
    num_output: 1024
    kernel_size: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}
layer {
  bottom: "fc9"
  top: "fc9"
  name: "relu9"
  type: "ReLU"
}
layer {
  bottom: "fc9"
  top: "fc9"
  name: "drop9"
  type: "Dropout"
  dropout_param {
    dropout_ratio: 0.5
  }
}

####

layer {
  bottom: "fc9"
  top: "fc10_sal"
  name: "fc10_sal"
  type: "Convolution"
  convolution_param {
    num_output: 2
    kernel_size: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}

layer {
  bottom: "fc10_sal"
  top: "fc10_interp"
  name: "fc10_interp"
  type: "Interp"
  interp_param {
    zoom_factor: 8
  }
}

layer {
  bottom: "fc10_interp"
  name: "fc10_sm"
  top: "fc10_sm"
  type: "Softmax"
}


layer {
  bottom: "junk"
  name: "silence"
  type: "Silence"
}