Help for using with Unity 3D about dlibdotnet HOT 5 CLOSED

You tried to use FrontalFaceDetector.Detect method for RGBA image?
If so, this method does not support RGBA image due to ristricts of dlib.

However, I may misunderstand what you faces to.
Would you show your code to recognize faces?

from dlibdotnet.

No, (i've tried this idead but, yes, the error message is "RGBAlphaPixel can not converted to Uint8")
What i desire to do is to detect the faces on resources online from the webcam and not from the images stocked on the machine. So instead of using LoadImage function (like used in the exemple); i use LoadImageData from a byte array to get Array2D. This byte array, I convert from the Color32 array provided by Unity.

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.IO;
using UnityEngine;
using DlibDotNet;
using System.Runtime.InteropServices;

namespace TestDlibDotNet
{
public class Class1
{
public int UnityTest(Color32[] color_array)
{
int lengt = 0;
using (var detector = FrontalFaceDetector.GetFrontalFaceDetector())
using (var img = Dlib.LoadImageData(Color32ArrayToByteArray(color_array), (uint)640, (uint)480, (uint)4))
{
Dlib.PyramidUp(img);
var dets = detector.Detect(img);

            try
            {
                lengt = dets.Length;
            }
            catch (InvalidCastException e)
            {
            }

        }
        return lengt;

    }
    public byte[] Color32ArrayToByteArray(Color32[] colors)
    {
        byte[] bytes = new byte[colors.Length * 4];
        for (int i = 0; i < bytes.Length/4; i += 4)
        {
            bytes[i]= colors[i].r;
            bytes[i+1] = colors[i].g;
            bytes[i+2] = colors[i].b;
            bytes[i+3] = colors[i].a;
           
        }
        return bytes;
    }


}

}

from dlibdotnet.

ok.
LoadImageData does not suit for this case.
LoadImageData treats argments as T type image. It means that this function does not convert color format.

If you expect this function convert image to grayscale, it's wrong.
In addition, 4th argment of LoadImageData is not channels but srtide.
In this case, It could be 640 x 4.

Therefore, you should convert image to grayscale or RGB manually.

from dlibdotnet.

Thanks, so finally, i need to manually convert my image resources from webcam to grayscale and then convert the last one to byte array ? What it means The 4th parameter (uint step) in LoadImageData ? :D

from dlibdotnet.

Step is Stride. This parameter equals to last parameter of System.Windows.Media.Imaging.BitmapSource.Create.
Generally, stride is byte size per line and it equals to image width * channels. However, some image format; eg bmp does not equals to.
Please get further information from https://en.wikipedia.org/wiki/BMP_file_format.

from dlibdotnet.