Python Base64 affect output data

Question

I am facing an issue with Base64 Encoder/Decoder I don't understand. I am transfering an image opened by :

image = open(IMAGE_PATH, "rb").read()
payload = {"image": image,
           "filename": os.path.basename(IMAGE_PATH),
           "extension": file_extension
r = requests.post(KERAS_REST_API_URL, files=payload).json()

once transfered, the image is opened by the REST API by :

image = Image.open(io.BytesIO(image))
image = np.asarray(image)

and then passed to Base64 Encoder => Everything is fine. See below the resul at the Base64 Encoder function

[[29477 29480 29443 ... 29465 29463 29473]
 [29473 29469 29440 ... 29481 29460 29441]
 [29434 29436 29452 ... 29459 29474 29444]
 ...
 [29455 29437 29440 ... 29439 29433 29481]
 [29451 29456 29441 ... 29469 29450 29452]
 [29472 29455 29440 ... 29433 29453 29459]]

[INFO] Image Shape : (768, 1024)
[INFO] Channel : 1
[INFO] Image Shape : (768, 1024)
[INFO] Image Shape Encode : (768, 1024)
[INFO] Image Shape Encode/Decode : (786432,)

Instead of transfering an image by "open(IMAGE_PATH, "rb").read()", I would like to transfert a Numpy Array as my image will also be dicom image.

So, I open the image file with the following commands :

image = np.asarray(Image.open(IMAGE_PATH))

and payload becoms :

payload = {"image": json.dumps(image, cls=NumpyArrayEncoder),
                       "filename": os.path.basename(IMAGE_PATH),
                       "extension": file_extension
                       }
requests.post line is unchanged.

In the REST API, the line that reads the image becomes :

image = request.files["image"].read()
image = json.loads(image)
image = np.asarray(image)

and then passed to Base64 Encoder without any modification of the code. I get the following result

[[29477 29480 29443 ... 29465 29463 29473]
 [29473 29469 29440 ... 29481 29460 29441]
 [29434 29436 29452 ... 29459 29474 29444]
 ...
 [29455 29437 29440 ... 29439 29433 29481]
 [29451 29456 29441 ... 29469 29450 29452]
 [29472 29455 29440 ... 29433 29453 29459]]

[INFO] Image Shape : (768, 1024)
[INFO] Channel : 1
[INFO] Image Shape : (768, 1024)
[INFO] Image Shape Encode : (768, 1024)
[INFO] Image Shape Encode/Decode : (1572864,)

Finally, the issue I have is when I decode the data, the shape doesn't match anymore with the expected shape (2 times higher).

Do you have any hint to solve this issue ?

Thank you

def base64_encode_image(a):
    # base64 encode the input NumPy array
    print("[INFO] Image Shape Encode : " + str(a.shape))
    a = base64.b64encode(a).decode("utf-8")
    a = bytes(a, encoding="utf-8")
    a = np.frombuffer(base64.decodebytes(a), dtype='int32')
    print("[INFO] Image Shape Encode/Decode : " + str(a.shape))
    return base64.b64encode(a).decode("utf-8")




def base64_decode_image(a, dtype, shape):
    if sys.version_info.major == 3:
        print("[INFO] Python 3")
        a = bytes(a, encoding="utf-8")
    # convert the string to a NumPy array using the supplied data
    # type and target shape
    a = np.frombuffer(base64.decodestring(a), dtype=dtype)
    print("[INFO] Image Shape Decode : " + str(a.shape))
    a = a.reshape(shape)
    # return the decoded image
    return a

Python Base64 affect output data

0 Answers