Dataset

This dataset was created by Ashifur Rahman

Contents

Overview

In Deepfake Detection Challenge, many public kernel use the faces at each frame. Here there are all the faces of the tran sample plitted into training set and validation set. Each set has faces from real and fake videos

Feel free the use this dataset to train your models in Deepfake Detection Challenge. If you find this dataset useful, please consider upvoting to it.

Creation method

The dataset was created by facenet_pytorch library. I used the following code to create the images from a video clip. A part of the code was taken from https://www.kaggle.com/timesler/facial-recognition-model-in-pytorch

` class DetectionPipeline: # Pipeline class for detecting faces in the frames of a video file.

def _init_(self, detector=None, n_frames=None, batch_size=60, resize=None):
  """Constructor for DetectionPipeline class.

  Keyword Arguments:
    n_frames {int} -- Total number of frames to load. These will be evenly spaced
      throughout the video. If not specified (i.e., None), all frames will be loaded.
      (default: {None})
    batch_size {int} -- Batch size to use with MTCNN face detector. (default: {32})
    resize {float} -- Fraction by which to resize frames from original prior to face
      detection. A value less than 1 results in downsampling and a value greater than
      1 result in upsampling. (default: {None})
  """
  self.detector = detector
  if detector is None:
    device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
    self.detector = MTCNN(image_size=256, margin=40, keep_all=True, factor=0.5, post_process=False, device=device).eval()
  self.n_frames = n_frames
  self.batch_size = batch_size
  self.resize = resize

def _call_(self, filename):
  """Load frames from an MP4 video and detect faces.

  Arguments:
    filename {str} -- Path to video.
  """
  torch.cuda.empty_cache()
  faces = []
  conf = []

  # Create video reader and find length
  v_cap = cv2.VideoCapture(filename)
  v_len = int(v_cap.get(cv2.CAP_PROP_FRAME_COUNT))

  # Pick 'n_frames' evenly spaced frames to sample
  if self.n_frames is None:
    sample = np.arange(0, v_len)
  else:
    sample = np.linspace(0, v_len - 1, self.n_frames).astype(int)

  # Loop through frames
  frames = []
  for j in range(v_len):
    success = v_cap.grab()
    if j in sample:
      # Load frame
      success, frame = v_cap.retrieve()
      if not success:
        continue
      frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
      frame = Image.fromarray(frame)

      # Resize frame to desired size
      if self.resize is not None:
        frame = frame.resize([int(d * self.resize) for d in frame.size])
      frames.append(frame)

      # When batch is full, detect faces and reset frame list
      if len(frames) % self.batch_size == 0 or j == sample[-1]:
        curr_faces, curr_conf = self.detector(frames, return_prob=True)
        faces.extend(curr_faces)
        conf.extend(curr_conf)
        frames = []

  v_cap.release()

  return faces, conf

`

` def save_images(detection_pipeline, filename, out_path): base_name = out_path + '/' + filename.split('/')[-1].split('.')[0] faces, conf = detection_pipeline(filename) for i in range(len(faces)): if faces[i] is None: continue for j in range(len(faces[i])): if conf[i][j] < 0.9: continue out_filename = base_name + '_' + str(i) + '_' + str(j) + '.png' out_img = faces[i][j].cpu().numpy() out_img = np.transpose(out_img, axes=[1, 2, 0]) out_img = cv2.cvtColor(out_img, cv2.COLOR_RGB2BGR) cv2.imwrite(out_filename, out_img)

`

This dataset contains a very small subset of dfdc dataset, with 2203 videos and related metadata

Metadata columns : Index(['filename', 'video.@index', 'video.@codec_time_base', 'video.@width', 'video.@height', 'video.@sample_aspect_ratio', 'video.@display_aspect_ratio', 'video.@level', 'video.@r_frame_rate', 'video.@time_base', 'video.@duration_ts', 'video.@duration', 'video.@bit_rate', 'video.@nb_frames', 'audio.@index', 'audio.@codec_name', 'audio.@codec_long_name', 'audio.@codec_time_base', 'audio.@duration_ts', 'audio.@duration', 'audio.@bit_rate', 'audio.@max_bit_rate', 'audio.@nb_frames', 'label', 'split', 'original', 'folder', 'wav.hash.cnt', 'original.cnt', 'md5', 'md5.orig', 'wav.hash', 'wav.hash.orig', 'pxl.hash', 'pxl.hash.orig'], dtype='object')

dfb-data/dfdc dataset hosted on Hugging Face and contributed by the HF Datasets community

A dataset for deepfake detection challenge.

dfb-data/dfdc-faces dataset hosted on Hugging Face and contributed by the HF Datasets community

laodeAlif/dfdc dataset hosted on Hugging Face and contributed by the HF Datasets community

Comparison of model performance on the DFDC dataset.

The DFDC dataset, which was specifically released for a corresponding Kaggle competition.

Content

Motion stabilized faces from the DFDC dataset

Please see this thread for more information on this dataset.

Overview

In Deepfake Detection Challenge, many public notebooks use the faces at equidistant frames of the videos. Here are all the faces of the test sample.

Feel free to use this dataset to test your models in Deepfake Detection Challenge. If you find this dataset useful, please consider upvoting it.

Creation Method

The videos were first used to extract keyframes(limited to Max 20 frames from each video) from them, and then the faces were detected and cropped from the extracted frames with the highest confidence (>= 95%) value.

The notebook used to create the image dataset from video samples can be found here

The DFDC dataset contains 100,000 images of faces manipulated using Deepfakes.

Dataset

This dataset was created by Wahab Arabo

Contents

Dataset

This dataset was created by Fakecatcher AI

Released under CC0: Public Domain

Contents

Comparison between different combinations of Mixformer. The results in the table are test with the DFDC dataset (in %).

Performance comparison (precision, recall, and F1-score) of deepfake detection models on DFDC and FaceForensics++ (FF++) datasets.

150x150 face images from some frames from every video in part 18 of Deepfake Detection train data set. Specifically 10 frames evenly taken from all parts of every video.

Ablation study: AUC and F1-score comparison on DFDC and FF++.

The global Drive Force Distribution Controller (DFDC) market is booming, projected to reach $7.2 billion by 2033 at a CAGR of 12%. Discover key market trends, leading players (Bosch, Continental, ZF), regional analysis, and growth drivers shaping this exciting automotive technology sector. Learn more about the impact of ADAS and electric vehicles on DFDC market growth.

A dataset for deepfake detection containing real and fake videos from multiple sources (DFDC, FF++, CelebDF-V2). There are two folders, the folder 'fake' contains the manipulated videos and the folder 'real' contains the original unmanipulated videos. 600 videos in each folder of which 200 comes from each of 3 datasets, and are taken at random.