In Direct3D 10.x and 11.x, you have both a texture resource and a shader resource view. You can get both back when you call DirectX::CreateDDSTextureFromFile() or DirectX::CreateWICTextureFromFile().

Microsoft::WRL::ComPtr<ID3D11ID3D11Resource> res;
Microsoft::WRL::ComPtr<ID3D11ShaderResourceView> srv;
HRESULT hr = DirectXCreateDDSTextureFromFile( d3dDevice.Get(), L"SEAFLOOR.DDS",
    &res, &srv )
ThrowIfFailed(hr);

// You use srv.Get() when you are rendering with SpriteBatch

UINT width, height;
GetTextureSize( res.Get(), &width, &height );


void GetTextureSize( ID3D11Resource* res, UINT* width, UINT* height )
{
    assert( res != 0 );

    // This is the most generic solution. you can make it a lot
    // simpler if you know it will always be a 2D texture file
    D3D11_RESOURCE_DIMENSION dim;
    res->GetType( &dim );
    switch(dim)
    {
        case D3D11_RESOURCE_DIMENSION_TEXTURE1D:
            { 
                 auto txt = reinterpret_cast<ID3D11Texture1D*>( res );
                 D3D11_TEXTURE1D_DESC desc;
                 txt->GetDesc( &txt );
                 if ( width ) *width = txt.Width;
                 if ( height ) *height = 1;
            }
            break;
        case D3D11_RESOURCE_DIMENSION_TEXTURE2D:
            {
                 auto txt = reinterpret_cast<ID3D11Texture2D*>( res );
                 D3D11_TEXTURE2D_DESC desc;
                 txt->GetDesc( &txt );
                 if ( width ) *width = txt.Width;
                 if ( height ) *height = txt.Height;
            }
            break;
        case D3D11_RESOURCE_DIMENSION_TEXTURE3D:
            {
                 auto txt = reinterpret_cast<ID3D11Texture3D*>( res );
                 D3D11_TEXTURE3D_DESC desc;
                 txt->GetDesc( &txt );
                 if ( width ) *width = txt.Width;
                 if ( height ) *height = txt.Height;
            }
            break;
        default:
             if ( width ) *width = 0;
             if ( height ) *height = 0;
            break;
    }
}

The 'resource' is essentially the object that holds the actual data (a vertex buffer, index buffer, a texture, etc.). The 'shader resource view' is the object that tells the rendering pipeline how to interpret the data. This split provides some additional capabilities. For example, you could have a single 'resource' which is an array of 6 2D textures. You can then have two different 'shader resource views' both pointing to the same 'resource': one is a srv that interprets it as 6 2D textures as an array, another is an srv that interprets it as one cubemap.

It does make the 'simple' case a little more complicated which is why you can just keep the srv around after loading a texture and not worry about the specific 'resource' most of the time.

http://msdn.microsoft.com/en-us/library/windows/desktop/ff476894(v=vs.85).aspx

Oh. i see... thanks for that, Didn't know exactly how SRVs worked and how to get the data out of them.
Sorry, still trying to get the hang of DX11 ...