Its done

Dockerfile

@@ -1,26 +1,11 @@
 FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04
 EXPOSE 7860
 
-ENV CONDA_DIR /opt/conda
-RUN apt-get update && \
-    apt-get install -y build-essential  && \
-    apt-get install -y wget && \
-    apt-get install -y libgl1-mesa-glx && \
-    apt-get install -y libglib2.0-0 && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/*
-
-RUN wget --quiet https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda.sh && \
-     /bin/bash ~/miniconda.sh -b -p /opt/conda
-
-ENV PATH=$CONDA_DIR/bin:$PATH
-
 ENV GRADIO_SERVER_NAME=0.0.0.0
 WORKDIR /workspace
 
-ADD environment.yaml /workspace/environment.yaml
-RUN conda env update -n base --file environment.yaml
-RUN pip install typing-extensions -U
+ADD requirements.txt /workspace/requirements.txt
+RUN pip install -r requirements.txt
 
 ADD app.py /workspace/
 CMD [ "python" , "/workspace/app.py" ]

app.py

@@ -1,13 +1,116 @@
-from fawkes.protection import Fawkes
+import cv2
 import gradio as gr
+import numpy as np
+from mtcnn_cv2 import MTCNN
 import os
 
-def predict(level, img):
-  # print(img)
-  fwks = Fawkes("extractor_2", '0', 1, mode=level)
-  fwks.run_protection([img], format='jpeg')
-  splt = img.split(".")
-  # print(os.listdir('/tmp'))
-  return splt[0] + "_cloaked." + splt[1]
+detector = MTCNN()
 
-gr.Interface(fn=predict, inputs=[gr.inputs.Dropdown(["low", "mid", "high"], label="Protection Level"), gr.inputs.Image(type='filepath')], outputs=gr.outputs.Image(type="pil")).launch()
+def predict(img, selection):
+  faces = detector.detect_faces(img)
+
+  privacy_fn = None
+  if(selection == "Low"):
+    opts = (anonymize_face_pixelate, 20)
+  elif(selection == "Medium"):
+    opts = (anonymize_face_pixelate, 10)
+  elif(selection == "High"):
+    opts = (anonymize_face_pixelate, 4)
+  elif(selection == "Emoji"):
+    opts = (anonymize_face_emoji, "smiley")
+  else:
+    raise Exception("I don't know how you did it but you chose something else.")
+
+  if len(faces) > 0:
+    for features in faces:
+      img = opts[0](img, features, opts[1])
+  else:
+    raise Exception("No faces detected");
+  return img
+
+def anonymize_face_pixelate(image, features, blocks=10):
+  bb = features['box']
+  face_crop = image[bb[1]:bb[1]+bb[3], bb[0]:bb[0]+bb[2]]
+  # Divide the input image into NxN blocks
+  (h,w) = face_crop.shape[:2]
+  xSteps = np.linspace(0, w, blocks + 1, dtype="int")
+  ySteps = np.linspace(0, h, blocks + 1, dtype="int")
+
+  # loop over the blocks in both x and y direction
+  for i in range(1, len(ySteps)):
+    for j in range(1, len(xSteps)):
+      # compute starting and ending (x, y)-coordinates
+      # for current block
+      startX = xSteps[j - 1]
+      startY = ySteps[i - 1]
+      endX = xSteps[j]
+      endY = ySteps[i]
+
+      # Extract the ROI using NumPy array slicing, compute the
+      # mean of the ROI, and then draw a rectangle with the
+      # mean RGB values over the ROI in teh original image
+      roi = face_crop[startY:endY, startX:endX]
+      (B, G, R) = [int(x) for x in cv2.mean(roi)[:3]]
+      cv2.rectangle(face_crop, (startX, startY), (endX, endY),
+                    (B,G,R), -1)
+
+  image[bb[1]:bb[1]+bb[3], bb[0]:bb[0]+bb[2]] = face_crop
+  return image
+
+def anonymize_face_emoji(img, features, name="smiley"):
+  bb = features['box']
+  (y, x) = (bb[1] + int(bb[3]/2), bb[0] + int(bb[2]/2))
+  (h,w) = (bb[3], bb[2])
+  # Get emoji with transparency
+  mask = cv2.imread('raccoon_emoji.png', -1)
+
+  mshape = max(h,w)
+  offset = int(mshape/2)
+
+  return overlay_transparent(img, mask, 
+                            x - offset,
+                            y - offset, 
+                            (mshape, mshape))
+
+def overlay_transparent(background_img, img_to_overlay_t, x, y, overlay_size=None):
+	"""
+	@brief      Overlays a transparant PNG onto another image using CV2
+	
+	@param      background_img    The background image
+	@param      img_to_overlay_t  The transparent image to overlay (has alpha channel)
+	@param      x                 x location to place the top-left corner of our overlay
+	@param      y                 y location to place the top-left corner of our overlay
+	@param      overlay_size      The size to scale our overlay to (tuple), no scaling if None
+	
+	@return     Background image with overlay on top
+	"""
+	
+	bg_img = background_img.copy()
+	
+	if overlay_size is not None:
+		img_to_overlay_t = cv2.resize(img_to_overlay_t.copy(), overlay_size)
+
+	# Extract the alpha mask of the RGBA image, convert to RGB 
+	b,g,r,a = cv2.split(img_to_overlay_t)
+	overlay_color = cv2.merge((b,g,r))
+	
+	# Apply some simple filtering to remove edge noise
+	mask = cv2.medianBlur(a,5)
+
+	h, w, _ = overlay_color.shape
+	roi = bg_img[y:y+h, x:x+w]
+
+	# Black-out the area behind the logo in our original ROI
+	img1_bg = cv2.bitwise_and(roi.copy(),roi.copy(),mask = cv2.bitwise_not(mask))
+	
+	# Mask out the logo from the logo image.
+	img2_fg = cv2.bitwise_and(overlay_color,overlay_color,mask = mask)
+
+	# Update the original image with our new ROI
+	bg_img[y:y+h, x:x+w] = cv2.add(img1_bg, img2_fg)
+
+	return bg_img
+
+gr.Interface(fn=predict,
+             inputs=[gr.components.Image(type='numpy'), gr.components.Radio(["Low", "Medium", "High", "Emoji"], value="Medium")],
+             outputs=gr.components.Image(type="pil")).launch(show_error=True)

environment.yaml

@@ -1,123 +0,0 @@
-name: gradio
-channels:
-  - nvidia
-  - conda-forge
-  - defaults
-dependencies:
-  - _libgcc_mutex=0.1=main
-  - _openmp_mutex=5.1=1_gnu
-  - ca-certificates=2022.9.24=ha878542_0
-  - certifi=2022.9.24=pyhd8ed1ab_0
-  - cudatoolkit=11.0.221=h6bb024c_0
-  - cudnn=8.0.4=cuda11.0_0
-  - libedit=3.1.20210910=h7f8727e_0
-  - libffi=3.2.1=hf484d3e_1007
-  - libgcc-ng=11.2.0=h1234567_1
-  - libgomp=11.2.0=h1234567_1
-  - libstdcxx-ng=11.2.0=h1234567_1
-  - ncurses=6.3=h5eee18b_3
-  - openssl=1.1.1o=h166bdaf_0
-  - pip=22.2.2=py38h06a4308_0
-  - python=3.8.0=h0371630_2
-  - readline=7.0=h7b6447c_5
-  - setuptools=63.4.1=py38h06a4308_0
-  - sqlite=3.33.0=h62c20be_0
-  - tk=8.6.12=h1ccaba5_0
-  - wheel=0.37.1=pyhd3eb1b0_0
-  - xz=5.2.6=h5eee18b_0
-  - zlib=1.2.12=h5eee18b_3
-  - pip:
-    - absl-py==0.15.0
-    - aiohttp==3.8.3
-    - aiosignal==1.2.0
-    - analytics-python==1.4.0
-    - anyio==3.6.1
-    - astunparse==1.6.3
-    - async-timeout==4.0.2
-    - attrs==22.1.0
-    - backoff==1.10.0
-    - bcrypt==4.0.1
-    - bleach==5.0.1
-    - cachetools==5.2.0
-    - cffi==1.15.1
-    - charset-normalizer==2.1.1
-    - click==8.1.3
-    - contourpy==1.0.5
-    - cryptography==38.0.1
-    - cycler==0.11.0
-    - fastapi==0.76.0
-    - fawkes==1.0.4
-    - ffmpy==0.3.0
-    - flatbuffers==1.12
-    - fonttools==4.37.4
-    - frozenlist==1.3.1
-    - fsspec==2022.8.2
-    - gast==0.3.3
-    - google-auth==2.12.0
-    - google-auth-oauthlib==0.4.6
-    - google-pasta==0.2.0
-    - gradio==3.0.20
-    - grpcio==1.32.0
-    - h11==0.14.0
-    - h5py==2.10.0
-    - idna==3.4
-    - importlib-metadata==5.0.0
-    - jinja2==3.1.2
-    - keras==2.4.3
-    - keras-preprocessing==1.1.2
-    - kiwisolver==1.4.4
-    - linkify-it-py==1.0.3
-    - markdown==3.4.1
-    - markdown-it-py==2.1.0
-    - markupsafe==2.1.1
-    - matplotlib==3.6.1
-    - mdit-py-plugins==0.3.1
-    - mdurl==0.1.2
-    - monotonic==1.6
-    - mtcnn==0.1.1
-    - multidict==6.0.2
-    - numpy==1.19.5
-    - oauthlib==3.2.1
-    - opencv-python==4.6.0.66
-    - opt-einsum==3.3.0
-    - orjson==3.8.0
-    - packaging==21.3
-    - pandas==1.4.4
-    - paramiko==2.11.0
-    - pillow==9.2.0
-    - protobuf==3.19.6
-    - pyasn1==0.4.8
-    - pyasn1-modules==0.2.8
-    - pycparser==2.21
-    - pycryptodome==3.15.0
-    - pydantic==1.9.2
-    - pydub==0.25.1
-    - pynacl==1.5.0
-    - pyparsing==3.0.9
-    - python-dateutil==2.8.2
-    - python-multipart==0.0.5
-    - pytz==2022.4
-    - pyyaml==6.0
-    - requests==2.28.1
-    - requests-oauthlib==1.3.1
-    - rsa==4.9
-    - scipy==1.9.2
-    - six==1.15.0
-    - sniffio==1.3.0
-    - starlette==0.18.0
-    - tensorboard==2.10.1
-    - tensorboard-data-server==0.6.1
-    - tensorboard-plugin-wit==1.8.1
-    - tensorflow==2.4.1
-    - tensorflow-estimator==2.4.0
-    - termcolor==1.1.0
-    - typing-extensions
-    - uc-micro-py==1.0.1
-    - urllib3==1.26.12
-    - uvicorn==0.18.3
-    - webencodings==0.5.1
-    - werkzeug==2.2.2
-    - wrapt==1.12.1
-    - yarl==1.8.1
-    - zipp==3.9.0
-prefix: /opt/conda/envs/gradio

requirements.txt

@@ -1 +0,0 @@
-fawkes==1.0.4

requirements_personal.txt

@@ -1,2 +0,0 @@
-gradio==3.0.20
-fawkes==1.0.4