scrabbled face in the video

[romasiugzdaite]

hi Enes,

we are trying to scramble just a face in the video, could you help us with that please?

[altunenes]

hello @romasiugzdaite !
I realized I didn't focus on video scramble as much as static images, but we can still use it. 🙂 (I think I need to update scramblery soon... 😃)

For now, to fix your issue, I tried to create a simple code by commenting as much as possible so I hope it will be a roadmap for the next update🙂... just run the code by changing the video input name, and it should work... (but bump me if you got any errors/questions)

Note: This is designed for a single face in a video, if you have to apply scramble for multiple faces in a video let me know because the code might change significantly... 🙂

from scramblery import scramblery
import cv2 ## use pip install opencv-python if you don't have this library.
def process_video(input_video_path, output_video_path, scramble_settings):
    cap = cv2.VideoCapture(input_video_path)

    # Checking if the video opened successfully....
    if not cap.isOpened():
        print("Error: Could not open video.")
        return

    frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = cap.get(cv2.CAP_PROP_FPS)
    codec = cv2.VideoWriter_fourcc(*'XVID')


    total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))

    out = cv2.VideoWriter(output_video_path, codec, fps, (frame_width, frame_height))

    frame_number = 0  # Keep track of frame number
    while cap.isOpened():
        ret, frame = cap.read()

        # if frame is read correctly, ret is True
        if not ret:
            print("Can't receive frame. Exiting ...")
            break

        frame_number += 1  # Increment frame number

        # Printing the progress...
        print(f"Processing frame {frame_number} of {total_frames} ({(frame_number/total_frames)*100:.2f}%)")

        # Try processing the frame through scrambleface function
        try:
            scrambled_frame = scramblery.scrambleface(frame, **scramble_settings)
        except TypeError:  # Catch the error when no faces are detected
            scrambled_frame = frame  # If an error occurs, use the original frame

        # Write the frame into the output video file
        out.write(scrambled_frame)

        # Uncomment this section if you want to see the video while it's being processed
        # cv2.imshow('frame', scrambled_frame)
        # if cv2.waitKey(1) == ord('q'):
        #     break

    # Release everything when the job is finished
    cap.release()
    out.release()
    cv2.destroyAllWindows()

# Usage example
input_video_path = 'enes.mp4'  # Update with your video in your path
output_video_path = 'enes_scramble.mp4'  # output video name

scramble_settings = {
    'splits': 25,  #scramble ratio
    'type': 'pixel', ## For now only "stack" and "pixel" are available. 
    'seamless': False, ## Seamless Cloning also needs to be update for next patch. so keep it as "false" for now.
    'bg': True, ## note If you set it to "false" the background disappears, but the background is still visible at every frame where the face is not detected. So I don't recommend it for now...
    'seed': None, 
    'write': False  # It's important to set 'write' to False so the function returns the image instead of saving it
}

process_video(input_video_path, output_video_path, scramble_settings)

[romasiugzdaite]

thanks, I will try now.

…

On Thu, Oct 19, 2023 at 1:50 PM Enes ***@***.***> wrote: hello @romasiugzdaite <https://github.com/romasiugzdaite> ! I realized I didn't focus on video scramble as much as static images, but we can still use it. 🙂 (I think I need to update scramblery soon... 😃) For now, to fix your issue, I tried to create a simple code by commenting as much as possible so I hope it will be a roadmap for the next update🙂... just run the code by changing the video input name, and it should work... (but bump me if you got any errors/questions) Note: This is designed for a single face in a video, if you have to apply scramble for multiple faces in a video let me know because the code might change significantly... 🙂 from scramblery import scrambleryimport cv2 ## use pip install opencv-python if you don't have this library.def process_video(input_video_path, output_video_path, scramble_settings): cap = cv2.VideoCapture(input_video_path) # Checking if the video opened successfully.... if not cap.isOpened(): print("Error: Could not open video.") return frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) fps = cap.get(cv2.CAP_PROP_FPS) codec = cv2.VideoWriter_fourcc(*'XVID') total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)) out = cv2.VideoWriter(output_video_path, codec, fps, (frame_width, frame_height)) frame_number = 0 # Keep track of frame number while cap.isOpened(): ret, frame = cap.read() # if frame is read correctly, ret is True if not ret: print("Can't receive frame. Exiting ...") break frame_number += 1 # Increment frame number # Printing the progress... print(f"Processing frame {frame_number} of {total_frames} ({(frame_number/total_frames)*100:.2f}%)") # Try processing the frame through scrambleface function try: scrambled_frame = scramblery.scrambleface(frame, **scramble_settings) except TypeError: # Catch the error when no faces are detected scrambled_frame = frame # If an error occurs, use the original frame # Write the frame into the output video file out.write(scrambled_frame) # Uncomment this section if you want to see the video while it's being processed # cv2.imshow('frame', scrambled_frame) # if cv2.waitKey(1) == ord('q'): # break # Release everything when the job is finished cap.release() out.release() cv2.destroyAllWindows() # Usage exampleinput_video_path = 'enes.mp4' # Update with your video in your pathoutput_video_path = 'enes_scramble.mp4' # output video name scramble_settings = { 'splits': 25, #scramble ratio 'type': 'pixel', ## For now only "stack" and "pixel" are available. 'seamless': False, ## Seamless Cloning also needs to be update for next patch. so keep it as "false" for now. 'bg': True, ## note If you set it to "false" the background disappears, but the background is still visible at every frame where the face is not detected. So I don't recommend it for now... 'seed': None, 'write': False # It's important to set 'write' to False so the function returns the image instead of saving it } process_video(input_video_path, output_video_path, scramble_settings) — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFHO7QLH3X6DO4AGIM42ATYAEHXVAVCNFSM6AAAAAA6GXVDA6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTONZQG4ZTQOJXGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[romasiugzdaite]

it worked!!!!!!!!! thank you Enes!

…

On Thu, Oct 19, 2023 at 1:50 PM Enes ***@***.***> wrote: hello @romasiugzdaite <https://github.com/romasiugzdaite> ! I realized I didn't focus on video scramble as much as static images, but we can still use it. 🙂 (I think I need to update scramblery soon... 😃) For now, to fix your issue, I tried to create a simple code by commenting as much as possible so I hope it will be a roadmap for the next update🙂... just run the code by changing the video input name, and it should work... (but bump me if you got any errors/questions) Note: This is designed for a single face in a video, if you have to apply scramble for multiple faces in a video let me know because the code might change significantly... 🙂 from scramblery import scrambleryimport cv2 ## use pip install opencv-python if you don't have this library.def process_video(input_video_path, output_video_path, scramble_settings): cap = cv2.VideoCapture(input_video_path) # Checking if the video opened successfully.... if not cap.isOpened(): print("Error: Could not open video.") return frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) fps = cap.get(cv2.CAP_PROP_FPS) codec = cv2.VideoWriter_fourcc(*'XVID') total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)) out = cv2.VideoWriter(output_video_path, codec, fps, (frame_width, frame_height)) frame_number = 0 # Keep track of frame number while cap.isOpened(): ret, frame = cap.read() # if frame is read correctly, ret is True if not ret: print("Can't receive frame. Exiting ...") break frame_number += 1 # Increment frame number # Printing the progress... print(f"Processing frame {frame_number} of {total_frames} ({(frame_number/total_frames)*100:.2f}%)") # Try processing the frame through scrambleface function try: scrambled_frame = scramblery.scrambleface(frame, **scramble_settings) except TypeError: # Catch the error when no faces are detected scrambled_frame = frame # If an error occurs, use the original frame # Write the frame into the output video file out.write(scrambled_frame) # Uncomment this section if you want to see the video while it's being processed # cv2.imshow('frame', scrambled_frame) # if cv2.waitKey(1) == ord('q'): # break # Release everything when the job is finished cap.release() out.release() cv2.destroyAllWindows() # Usage exampleinput_video_path = 'enes.mp4' # Update with your video in your pathoutput_video_path = 'enes_scramble.mp4' # output video name scramble_settings = { 'splits': 25, #scramble ratio 'type': 'pixel', ## For now only "stack" and "pixel" are available. 'seamless': False, ## Seamless Cloning also needs to be update for next patch. so keep it as "false" for now. 'bg': True, ## note If you set it to "false" the background disappears, but the background is still visible at every frame where the face is not detected. So I don't recommend it for now... 'seed': None, 'write': False # It's important to set 'write' to False so the function returns the image instead of saving it } process_video(input_video_path, output_video_path, scramble_settings) — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFHO7QLH3X6DO4AGIM42ATYAEHXVAVCNFSM6AAAAAA6GXVDA6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTONZQG4ZTQOJXGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[altunenes]

glad that worked! 🙂 I will extend this project soon to prevent these kinds of problems...

[romasiugzdaite]

Amazing! I will advertise you on twitter!

…

On Thu, 19 Oct 2023, 17:08 Enes, ***@***.***> wrote: glad that worked! 🙂 I will extend this project soon to prevent these kinds of problems... — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFHO7SQ5A746VFMRBZ62ZTYAE7AFAVCNFSM6AAAAAA6GXVDA6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTONZRGE4DKNZUGY> . You are receiving this because you were mentioned.Message ID: ***@***.***>