Image transfer through socket between C++ and C# application

Question

I have server application in C++ that making screenshot and it must be transferred to client, that written in C#. I've encountered few problems:

1. I can't just transfer bmp image because its too heavy, so i need to convert image to another format (jpg, png maybe?)
2. I need to retreive picture's byte array on client side and use it as picturebox's image.

The problem is that I don't know proper algorithm of achieving this. Saving image to file on server side and transferring a file and then saving file on client and only then reading it to picturebox is not a solution either. So, summarizing my task: How can i convert bitmap to light-weight format and transfer it to client?

Thank you for your time!

Answer 1

Much thanks for answering that question!After some research on stack I found similar questions, and there was a code that converts jpg image to byte array that is ready to be transferred through socket. Here it is (some variables may be unused)

#include <objidl.h>
#include <Gdiplus.h>
#pragma comment(lib, "gdiplus.lib")

int GetEncoderClsid(WCHAR *format, CLSID *pClsid)
{
    unsigned int num = 0, size = 0;
    GetImageEncodersSize(&num, &size);
    if (size == 0) return -1;
    ImageCodecInfo *pImageCodecInfo = (ImageCodecInfo *)(malloc(size));
    if (pImageCodecInfo == NULL) return -1;
    GetImageEncoders(num, size, pImageCodecInfo);
    for (unsigned int j = 0; j < num; ++j)
    {
        if (wcscmp(pImageCodecInfo[j].MimeType, format) == 0){
            *pClsid = pImageCodecInfo[j].Clsid;
            free(pImageCodecInfo);
            return j;
        }
    }
    free(pImageCodecInfo);
    return -1;
}

BYTE *GetScreeny(LPWSTR lpszFilename, ULONG uQuality, int *buff_size) // by Napalm
{
    ULONG_PTR gdiplusToken;
    GdiplusStartupInput gdiplusStartupInput;
    GdiplusStartup(&gdiplusToken, &gdiplusStartupInput, NULL);
    HWND hMyWnd = GetDesktopWindow(); // get my own window
    RECT  r;             // the area we are going to capture 
    int w, h;            // the width and height of the area
    HDC dc;              // the container for the area
    int nBPP;
    HDC hdcCapture;
    LPBYTE lpCapture;
    int nCapture;
    int iRes;
    CLSID imageCLSID;
    Bitmap *pScreenShot;
    HGLOBAL hMem;
    int result;

    // get the area of my application's window      
    //GetClientRect(hMyWnd, &r);
    GetWindowRect(hMyWnd, &r);
    dc = GetWindowDC(hMyWnd);//   GetDC(hMyWnd) ;
    w = r.right - r.left;
    h = r.bottom - r.top;
    nBPP = GetDeviceCaps(dc, BITSPIXEL);
    hdcCapture = CreateCompatibleDC(dc);


    // create the buffer for the screenshot
    BITMAPINFO bmiCapture = {
        sizeof(BITMAPINFOHEADER), w, -h, 1, nBPP, BI_RGB, 0, 0, 0, 0, 0,
    };

    // create a container and take the screenshot
    HBITMAP hbmCapture = CreateDIBSection(dc, &bmiCapture,
        DIB_PAL_COLORS, (LPVOID *)&lpCapture, NULL, 0);

    // failed to take it
    if (!hbmCapture)
    {
        DeleteDC(hdcCapture);
        DeleteDC(dc);
        GdiplusShutdown(gdiplusToken);
        printf("failed to take the screenshot. err: %d\n", GetLastError());
        return 0;
    }

    // copy the screenshot buffer
    nCapture = SaveDC(hdcCapture);
    SelectObject(hdcCapture, hbmCapture);
    BitBlt(hdcCapture, 0, 0, w, h, dc, 0, 0, SRCCOPY);
    RestoreDC(hdcCapture, nCapture);
    DeleteDC(hdcCapture);
    DeleteDC(dc);

    GpImage *bob;
    IStream *ssStr;

    // save the buffer to a file    
    pScreenShot = new Bitmap(hbmCapture, (HPALETTE)NULL);
    EncoderParameters encoderParams;
    encoderParams.Count = 1;
    encoderParams.Parameter[0].NumberOfValues = 1;
    encoderParams.Parameter[0].Guid = EncoderQuality;
    encoderParams.Parameter[0].Type = EncoderParameterValueTypeLong;
    encoderParams.Parameter[0].Value = &uQuality;
    GetEncoderClsid(L"image/jpeg", &imageCLSID);

    IStream *pStream = NULL;
    LARGE_INTEGER liZero = {};
    ULARGE_INTEGER pos = {};
    STATSTG stg = {};
    ULONG bytesRead = 0;
    HRESULT hrRet = S_OK;

    //BYTE* buffer = NULL;  // this is your buffer that will hold the jpeg bytes
    DWORD dwBufferSize = 0;  // this is the size of that buffer;


    hrRet = CreateStreamOnHGlobal(NULL, TRUE, &pStream);
    hrRet = pScreenShot->Save(pStream, &imageCLSID, &encoderParams) == 0 ? S_OK : E_FAIL;
    hrRet = pStream->Seek(liZero, STREAM_SEEK_SET, &pos);
    hrRet = pStream->Stat(&stg, STATFLAG_NONAME);

    // allocate a byte buffer big enough to hold the jpeg stream in memory
    BYTE *buffer = new BYTE[stg.cbSize.LowPart];
    //buffer = (BYTE *)malloc(stg.cbSize.LowPart);
    hrRet = (buffer == NULL) ? E_OUTOFMEMORY : S_OK;
    dwBufferSize = stg.cbSize.LowPart;
    //wchar_t message[512];
    //wsprintf(message, L"%d", stg.cbSize.LowPart);
    //MessageBox(NULL, message, NULL, MB_OK);
    // copy the stream into memory
    hrRet = pStream->Read(buffer, stg.cbSize.LowPart, &bytesRead);

    // now go save "buffer" and "dwBufferSize" off somewhere.  This is the jpeg buffer
    // don't forget to free it when you are done

    // After success or if any of the above calls fail, don't forget to release the stream
    if (pStream)
    {
        pStream->Release();
    }

    delete pScreenShot;
    DeleteObject(hbmCapture);
    GdiplusShutdown(gdiplusToken);
    *buff_size = (int)dwBufferSize;
    //return iRes;
    return buffer;

}

Answer 2

If you can use .Net managed code from your c++ application then you can use native libs (see https://stackoverflow.com/a/3517974/149818).

If you are restricted by clear c++ only please see LodePNG

Answer 3

You can use libpng at c++ side for convertion to .png. At C# side Bitmap class .png can be loaded by just:

new Bitmap(png);

Remember: I may be wrong about C# side.