Skip to content

DCTHidePNGinWAV

stegobox.codec.DCTHidePNGinWAV

This module implements simple DCT steganography in WAV audios.

the code by zhudehua

Source code in stegobox/codec/dct_hide_pnginwav.py 
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
class DCTHidePNGinWAV:
    """
    This module implements simple DCT steganography in WAV audios.

    the code by zhudehua
    """

    def __init__(self) -> None:
        pass

    def encode(
        self, carrier: wave.Wave_read, payload: np.ndarray
    ) -> tuple[bytes, wave._wave_params, tuple[int, int, int]]:
        """Encoder requires the carrier audio to be WAV and the payload to be a image.

        Args:
            carrier: Carrier audio in format WAV. Read with `stegobox.io.audio.read()`.
            payload: Payload (secret image) to be encoded. Read with
                `stegobox.io.image.read_cv2()`.

        Returns:
            Frame information in bytes of the audio with the payload embedded.
            Parameters for audio with the payload embedded.
            the size of secret image.
        """
        # make string to bit
        image_size = payload.shape
        payload = payload.reshape(-1)
        message_bits = np.unpackbits(payload)
        frame_num = carrier.getnframes()
        wave_head = carrier.getparams()
        frame_rate = carrier.getframerate()
        frame_data = carrier.readframes(frame_num)
        # calculate the length of dct data
        bin_data = bytearray(frame_data)
        block_size = len(bin_data) // 8
        zong_length = block_size * 8
        dct_data = bin_data[:zong_length]
        print("block num is ", block_size)
        print("num is ", len(message_bits))
        if len(message_bits) > block_size:
            print("the size of image is too large!")
            raise Exception("the image size is larger than the audio size")

        # deal the left data
        left_data = bin_data[zong_length:]
        left = []
        for i in range(len(left_data)):
            left.append(left_data[i])
            left.append(left_data[i])
        # make the carrier data to N*4*2 block for dct's input
        dct_data = np.array(dct_data).reshape(-1, 2)
        dct_data = dct_data.astype(np.float32)
        dct_data = np.vsplit(dct_data, block_size)
        temp_dct = np.zeros(shape=(block_size, 4, 2))
        # DCT
        for i in range(len(dct_data)):
            temp_dct[i] = cv2.dct(dct_data[i])

        out_audio_dct = []
        end = 0
        #   insert into secret info
        for i in range(len(message_bits)):
            end = end + 1
            out_audio_dct.append(temp_dct[i])
            te = temp_dct[i].copy()
            if message_bits[i] == 0:
                te[0][0] = te[0][0] - 10
            out_audio_dct.append(te)

        for j in range(end + 1, len(temp_dct)):
            out_audio_dct.append(temp_dct[j])
            out_audio_dct.append(temp_dct[j])

        outdata = np.zeros(shape=(2 * block_size, 4, 2))

        for i in range(len(out_audio_dct)):
            outdata[i] = cv2.idct(out_audio_dct[i])
        outdata = np.trunc(outdata).astype(np.uint8)
        outdata = outdata.reshape(-1)
        outdata = bytes(outdata)
        outdata = outdata + bytes(left)
        wave_head = wave_head._replace(framerate=2 * frame_rate)
        wave_head = wave_head._replace(nframes=2 * frame_num)
        return outdata, wave_head, (image_size[0], image_size[1], image_size[2])

    def decode(
        self, carrier: wave.Wave_read, image_size: tuple[int, int, int]
    ) -> np.ndarray:

        """Decode the secret payload from the carrier audio

        Args:
            carrier: Carrier audio in format WAV. Read with `stegobox.io.audio.read()`.
            image_size: the size of secret iamge.in format [width,hight,channels]

        Returns:
            The decoded payload (secret image).
        """
        image_size = image_size
        pay_length = image_size[0] * image_size[1] * image_size[2] * 8
        frame_data = carrier.readframes(carrier.getnframes())
        bin_data = bytearray(frame_data)
        # divide into  4*2 block
        block_size = len(bin_data) // 8
        zong_length = block_size * 8
        dct_data = bin_data[:zong_length]
        dct_data = np.array(dct_data).reshape(-1, 2)
        dct_data = dct_data.astype(np.float32)
        dct_data = np.vsplit(dct_data, block_size)
        # used to save data after dct
        temp_dct = np.zeros(shape=(block_size, 4, 2))
        for i in range(len(dct_data)):
            temp_dct[i] = cv2.dct(dct_data[i])

        # used to save extract data in format bit
        outdata = []
        for i in range(0, 2 * pay_length, 2):
            if temp_dct[i][0][0] == temp_dct[i + 1][0][0]:
                outdata.append(1)
            else:
                outdata.append(0)
        outdata = np.packbits(outdata)
        outdata = outdata.reshape(image_size)

        return outdata

encode(carrier, payload)

Encoder requires the carrier audio to be WAV and the payload to be a image.

Parameters:

Name Type Description Default
carrier Wave_read

Carrier audio in format WAV. Read with stegobox.io.audio.read().

 required
payload ndarray

Payload (secret image) to be encoded. Read with stegobox.io.image.read_cv2().

 required

Returns:

Type Description
bytes

Frame information in bytes of the audio with the payload embedded.
_wave_params

Parameters for audio with the payload embedded.
tuple[int, int, int]

the size of secret image.
Source code in stegobox/codec/dct_hide_pnginwav.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
def encode(
    self, carrier: wave.Wave_read, payload: np.ndarray
) -> tuple[bytes, wave._wave_params, tuple[int, int, int]]:
    """Encoder requires the carrier audio to be WAV and the payload to be a image.

    Args:
        carrier: Carrier audio in format WAV. Read with `stegobox.io.audio.read()`.
        payload: Payload (secret image) to be encoded. Read with
            `stegobox.io.image.read_cv2()`.

    Returns:
        Frame information in bytes of the audio with the payload embedded.
        Parameters for audio with the payload embedded.
        the size of secret image.
    """
    # make string to bit
    image_size = payload.shape
    payload = payload.reshape(-1)
    message_bits = np.unpackbits(payload)
    frame_num = carrier.getnframes()
    wave_head = carrier.getparams()
    frame_rate = carrier.getframerate()
    frame_data = carrier.readframes(frame_num)
    # calculate the length of dct data
    bin_data = bytearray(frame_data)
    block_size = len(bin_data) // 8
    zong_length = block_size * 8
    dct_data = bin_data[:zong_length]
    print("block num is ", block_size)
    print("num is ", len(message_bits))
    if len(message_bits) > block_size:
        print("the size of image is too large!")
        raise Exception("the image size is larger than the audio size")

    # deal the left data
    left_data = bin_data[zong_length:]
    left = []
    for i in range(len(left_data)):
        left.append(left_data[i])
        left.append(left_data[i])
    # make the carrier data to N*4*2 block for dct's input
    dct_data = np.array(dct_data).reshape(-1, 2)
    dct_data = dct_data.astype(np.float32)
    dct_data = np.vsplit(dct_data, block_size)
    temp_dct = np.zeros(shape=(block_size, 4, 2))
    # DCT
    for i in range(len(dct_data)):
        temp_dct[i] = cv2.dct(dct_data[i])

    out_audio_dct = []
    end = 0
    #   insert into secret info
    for i in range(len(message_bits)):
        end = end + 1
        out_audio_dct.append(temp_dct[i])
        te = temp_dct[i].copy()
        if message_bits[i] == 0:
            te[0][0] = te[0][0] - 10
        out_audio_dct.append(te)

    for j in range(end + 1, len(temp_dct)):
        out_audio_dct.append(temp_dct[j])
        out_audio_dct.append(temp_dct[j])

    outdata = np.zeros(shape=(2 * block_size, 4, 2))

    for i in range(len(out_audio_dct)):
        outdata[i] = cv2.idct(out_audio_dct[i])
    outdata = np.trunc(outdata).astype(np.uint8)
    outdata = outdata.reshape(-1)
    outdata = bytes(outdata)
    outdata = outdata + bytes(left)
    wave_head = wave_head._replace(framerate=2 * frame_rate)
    wave_head = wave_head._replace(nframes=2 * frame_num)
    return outdata, wave_head, (image_size[0], image_size[1], image_size[2])

decode(carrier, image_size)

Decode the secret payload from the carrier audio

Parameters:

Name Type Description Default
carrier Wave_read

Carrier audio in format WAV. Read with stegobox.io.audio.read().

 required
image_size tuple[int, int, int]

the size of secret iamge.in format [width,hight,channels]

 required

Returns:

Type Description
ndarray

The decoded payload (secret image).
Source code in stegobox/codec/dct_hide_pnginwav.py 
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
def decode(
    self, carrier: wave.Wave_read, image_size: tuple[int, int, int]
) -> np.ndarray:

    """Decode the secret payload from the carrier audio

    Args:
        carrier: Carrier audio in format WAV. Read with `stegobox.io.audio.read()`.
        image_size: the size of secret iamge.in format [width,hight,channels]

    Returns:
        The decoded payload (secret image).
    """
    image_size = image_size
    pay_length = image_size[0] * image_size[1] * image_size[2] * 8
    frame_data = carrier.readframes(carrier.getnframes())
    bin_data = bytearray(frame_data)
    # divide into  4*2 block
    block_size = len(bin_data) // 8
    zong_length = block_size * 8
    dct_data = bin_data[:zong_length]
    dct_data = np.array(dct_data).reshape(-1, 2)
    dct_data = dct_data.astype(np.float32)
    dct_data = np.vsplit(dct_data, block_size)
    # used to save data after dct
    temp_dct = np.zeros(shape=(block_size, 4, 2))
    for i in range(len(dct_data)):
        temp_dct[i] = cv2.dct(dct_data[i])

    # used to save extract data in format bit
    outdata = []
    for i in range(0, 2 * pay_length, 2):
        if temp_dct[i][0][0] == temp_dct[i + 1][0][0]:
            outdata.append(1)
        else:
            outdata.append(0)
    outdata = np.packbits(outdata)
    outdata = outdata.reshape(image_size)

    return outdata