| import numpy as np
|
| import torch
|
| import cv2
|
| import os
|
| import time
|
|
|
|
|
| class RivaWatermark(object):
|
| encoder = None
|
| decoder = None
|
|
|
| def __init__(self, watermarks=[], wmLen=32, threshold=0.52):
|
| self._watermarks = watermarks
|
| self._threshold = threshold
|
| if wmLen not in [32]:
|
| raise RuntimeError('rivaGan only supports 32 bits watermarks now.')
|
| self._data = torch.from_numpy(np.array([self._watermarks], dtype=np.float32))
|
|
|
| @classmethod
|
| def loadModel(cls):
|
| try:
|
| import onnxruntime
|
| except ImportError:
|
| raise ImportError(
|
| "The `RivaWatermark` class requires onnxruntime to be installed. "
|
| "You can install it with pip: `pip install onnxruntime`."
|
| )
|
|
|
| if RivaWatermark.encoder and RivaWatermark.decoder:
|
| return
|
| modelDir = os.path.dirname(os.path.abspath(__file__))
|
| RivaWatermark.encoder = onnxruntime.InferenceSession(
|
| os.path.join(modelDir, 'rivagan_encoder.onnx'))
|
| RivaWatermark.decoder = onnxruntime.InferenceSession(
|
| os.path.join(modelDir, 'rivagan_decoder.onnx'))
|
|
|
| def encode(self, frame):
|
| if not RivaWatermark.encoder:
|
| raise RuntimeError('call loadModel method first')
|
|
|
| frame = torch.from_numpy(np.array([frame], dtype=np.float32)) / 127.5 - 1.0
|
| frame = frame.permute(3, 0, 1, 2).unsqueeze(0)
|
|
|
| inputs = {
|
| 'frame': frame.detach().cpu().numpy(),
|
| 'data': self._data.detach().cpu().numpy()
|
| }
|
|
|
| outputs = RivaWatermark.encoder.run(None, inputs)
|
| wm_frame = outputs[0]
|
| wm_frame = torch.clamp(torch.from_numpy(wm_frame), min=-1.0, max=1.0)
|
| wm_frame = (
|
| (wm_frame[0, :, 0, :, :].permute(1, 2, 0) + 1.0) * 127.5
|
| ).detach().cpu().numpy().astype('uint8')
|
|
|
| return wm_frame
|
|
|
| def decode(self, frame):
|
| if not RivaWatermark.decoder:
|
| raise RuntimeError('you need load model first')
|
|
|
| frame = torch.from_numpy(np.array([frame], dtype=np.float32)) / 127.5 - 1.0
|
| frame = frame.permute(3, 0, 1, 2).unsqueeze(0)
|
| inputs = {
|
| 'frame': frame.detach().cpu().numpy(),
|
| }
|
| outputs = RivaWatermark.decoder.run(None, inputs)
|
| data = outputs[0][0]
|
| return np.array(data > self._threshold, dtype=np.uint8)
|
|
|
