Python 使用Keras.NET进行阵列整形

我无法将重塑表达式从R转换为C

这是R代码：

load_image <- function(file, ...) {
  magick::image_read(file, ...) %>%
    magick::image_data('rgba') %>%
    as.numeric()
}

img <- load_image("./my_image.png")
print(cat("Image length: ", length(img)))
print(cat("Image dim: ", dim(img)))

data <- array(dim = c(1, dim(img)))
data[1,,,] <- img

print(cat("Data length: ", length(data)))
print(cat("Data dim: ", dim(data)))

在我的C#代码中，我有：

var imgBytes = File.ReadAllBytes("./my_image.png");

using (var pngImage = new MagickImage(imgBytes)) {
    var img = np.array<ushort>(pngImage.GetPixels().ToArray());
    
    Console.WriteLine("Image length: {0}", img.len);
    Console.WriteLine("Image dim: {0}", string.join(" ", img.shape.dimensions));

    var data = img.reshape(1);

    //model.Predict(data);
}

如何以C#与R中相同的方式加载和重塑图像

var imgBytes = File.ReadAllBytes("./my_image.png");

using (var pngImage = new MagickImage(imgBytes)) {
    var img = np.array<ushort>(pngImage.GetPixels().ToArray());
    
    Console.WriteLine("Image length: {0}", img.len);
    Console.WriteLine("Image dim: {0}", string.join(" ", img.shape.dimensions));

    var data = img.reshape(1);

    //model.Predict(data);
}

Image length: 36000
Image dim: 36000