inline void im2col_nd_core_cpu(const Dtype* data_input, const bool im2col,
const int num_spatial_axes, const int* im_shape, const int* col_shape,
const int* kernel_shape, const int* pad, const int* stride,
const int* dilation, Dtype* data_output) {
if (!im2col) {
int im_size = im_shape[0];
for (int i = 0; i < num_spatial_axes; ++i) {
im_size *= im_shape[1 + i];
}
caffe_set(im_size, Dtype(0), data_output);
}
int kernel_size = 1;
for (int i = 0; i < num_spatial_axes; ++i) {
kernel_size *= kernel_shape[i];
}
const int channels_col = col_shape[0];
vector<int> d_offset(num_spatial_axes, 0);
vector<int> d_iter(num_spatial_axes, 0);
for (int c_col = 0; c_col < channels_col; ++c_col) {
// Loop over spatial axes in reverse order to compute a per-axis offset.
int offset = c_col;
for (int d_i = num_spatial_axes - 1; d_i >= 0; --d_i) {
if (d_i < num_spatial_axes - 1) {
offset /= kernel_shape[d_i + 1];
}
d_offset[d_i] = offset % kernel_shape[d_i];
}
for (bool incremented = true; incremented; ) {
// Loop over spatial axes in forward order to compute the indices in the
// image and column, and whether the index lies in the padding.
int index_col = c_col;
int index_im = c_col / kernel_size;
bool is_padding = false;
for (int d_i = 0; d_i < num_spatial_axes; ++d_i) {
const int d = d_iter[d_i];
const int d_im = d * stride[d_i] - pad[d_i] +
d_offset[d_i] * dilation[d_i];
is_padding |= d_im < 0 || d_im >= im_shape[d_i + 1];
index_col *= col_shape[d_i + 1];
index_col += d;
index_im *= im_shape[d_i + 1];
index_im += d_im;
}
if (im2col) {
if (is_padding) {
data_output[index_col] = 0;
} else {
data_output[index_col] = data_input[index_im];
}
} else if (!is_padding) { // col2im
data_output[index_im] += data_input[index_col];
}
// Loop over spatial axes in reverse order to choose an index,
// like counting.
incremented = false;
for (int d_i = num_spatial_axes - 1; d_i >= 0; --d_i) {
const int d_max = col_shape[d_i + 1];
DCHECK_LT(d_iter[d_i], d_max);
if (d_iter[d_i] == d_max - 1) {
d_iter[d_i] = 0;
} else { // d_iter[d_i] < d_max - 1
++d_iter[d_i];
incremented = true;
break;
}
}
} // while(incremented) {
} // for (int c = 0; c < channels_col; ++c) {
}
inline void im2col_nd_core_cpu(const Dtype* data_input, const bool im2col,
const int num_spatial_axes, const int* im_shape, const int* col_shape,
const int* kernel_shape, const int* pad, const int* stride,
const int* dilation, Dtype* data_output)
{
// 如果不是im2col则表明是col2im,也就是说data_output是需要输出的原始图像大小的数据
if (!im2col)
{
int im_size = im_shape[0];
for (int i = 0; i < num_spatial_axes; ++i)
{
im_size *= im_shape[1 + i];
}
caffe_set(im_size, Dtype(0), data_output);
}
// 一个kernel大小的块有多大
int kernel_size = 1;
for (int i = 0; i < num_spatial_axes; ++i)
{
kernel_size *= kernel_shape[i];
}
// channels_col = inputchannel(输入图像的channel)*kernel_size
const int channels_col = col_shape[0];
// 类似于im2col中的w_offset和h_offset，只不过因为这里是n维，所以用数组表示
vector<int> d_offset(num_spatial_axes, 0);
// 类似于im2col中w和h，是col_buff中的偏移
vector<int> d_iter(num_spatial_axes, 0);
for (int c_col = 0; c_col < channels_col; ++c_col)
{
// Loop over spatial axes in reverse order to compute a per-axis offset.
// Loop over spatial axes in reverse order to compute a per-axis offset.
// 计算n维kernel上的offset,与im2col中对应的代码一样的道理
// 只不过这里是n维了，所以用d_offset来表示
// 注意，这里用逆序来进行计算得到每个轴的偏移
int offset = c_col;
for (int d_i = num_spatial_axes - 1; d_i >= 0; --d_i)
{
if (d_i < num_spatial_axes - 1)
{
offset /= kernel_shape[d_i + 1];
}
d_offset[d_i] = offset % kernel_shape[d_i];
}
for (bool incremented = true; incremented; )
{
// Loop over spatial axes in forward order to compute the indices in the
// image and column, and whether the index lies in the padding.
// 是经过im2colnd变换之后的索引
int index_col = c_col;
// index_im是原始图像中的channel
int index_im = c_col / kernel_size;
bool is_padding = false;
for (int d_i = 0; d_i < num_spatial_axes; ++d_i)
{
// d是col_buff上的偏移，与d_pad相对(d_pad是原始图像上的偏移)
const int d = d_iter[d_i];
// 在d_pad是经过pad之后的col_buff中的坐标经过转换成原图中的坐标
const int d_im = d * stride[d_i] - pad[d_i] +
d_offset[d_i] * dilation[d_i];
// 判断经过im2colnd处理的图像上的像素是否位于输入的n维图像的上的pad的那个部分
is_padding |= d_im < 0 || d_im >= im_shape[d_i + 1];
// 计算位于col_buff中的位置(就是经过im2colnd变换之后的)
index_col *= col_shape[d_i + 1];
index_col += d;
// 计算位于原始图像中的位置
index_im *= im_shape[d_i + 1];
index_im += d_im;
}
if (im2col)
{
if (is_padding)
{
// 如果是位于pad的部分则设置为0
data_output[index_col] = 0;
}
else
{
data_output[index_col] = data_input[index_im];
}
}
else if (!is_padding)
{
// col2im
data_output[index_im] += data_input[index_col];
}
// 更新位于col_buff上的偏移d(d_iter就是所有的d存进去的)
// Loop over spatial axes in reverse order to choose an index,
// like counting.
incremented = false;
for (int d_i = num_spatial_axes - 1; d_i >= 0; --d_i)
{
const int d_max = col_shape[d_i + 1];
DCHECK_LT(d_iter[d_i], d_max);
if (d_iter[d_i] == d_max - 1)
{
d_iter[d_i] = 0;
}
else
{
// d_iter[d_i] < d_max - 1
++d_iter[d_i];
incremented = true;
break;
}
}
} // while(incremented) {
} // for (int c = 0; c < channels_col; ++c) {
}