为CNTK中的Resnet18优化的图像调整大小

Question

我试图使用在CNTK（C＃）中在Imagenet上训练的Resnet18预测图像类。为此，我必须调整输入图像的大小以适合CNN的输入要求。我当前的脚本是：

public static Bitmap CustomResize(Bitmap image, int width, int height, bool useHighQuality)
    {
        var newImg = new Bitmap(width, height);

        newImg.SetResolution(image.HorizontalResolution, image.VerticalResolution);

        using (var g = Graphics.FromImage(newImg))
        {
            g.CompositingMode = System.Drawing.Drawing2D.CompositingMode.SourceCopy;
            if (useHighQuality)
            {
                g.InterpolationMode = System.Drawing.Drawing2D.InterpolationMode.HighQualityBicubic;
                g.CompositingQuality = System.Drawing.Drawing2D.CompositingQuality.HighQuality;
                g.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.HighQuality;
                g.PixelOffsetMode = System.Drawing.Drawing2D.PixelOffsetMode.HighQuality;
            }
            else
            {
                g.InterpolationMode = System.Drawing.Drawing2D.InterpolationMode.Default;
                g.CompositingQuality = System.Drawing.Drawing2D.CompositingQuality.Default;
                g.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.Default;
                g.PixelOffsetMode = System.Drawing.Drawing2D.PixelOffsetMode.Default;
            }

            var attributes = new ImageAttributes();
            attributes.SetWrapMode(System.Drawing.Drawing2D.WrapMode.TileFlipXY);
            g.DrawImage(image, new System.Drawing.Rectangle(0, 0, width, height), 0, 0, image.Width, image.Height, GraphicsUnit.Pixel, attributes);
        }

        return newImg;
    }

但这不是优化的例程。在Python中，此任务很简单。但是我找不到C＃的同等优化版本。如果这是一个常见问题，我希望这里的许多人以前都已经解决了。