transparent-background 1.3.3

Make images with transparent background

Transparent Background

This is a background removing tool powered by InSPyReNet (ACCV 2022). You can easily remove background from the image or video or bunch of other stuffs when you can make the background transparent!

Image	Video	Webcam

:newspaper: News

• Our package is currently not working properly on small images without --fast argument. Sorry for the inconvenience and we'll fix this issue with better algorithm coming out shortly.
• [2023.09.22] For the issue with small images without --fast argument, please download This Checkpoint. After some user feedback (create issue or contact me), I'll decide to substitute the current checkpoint to the newer one or train again with different approach.
• [2023.09.25] The above checkpoint is now available with --mode base-nightly argument. --fast argument is deprecated. Use --mode [MODE] instead. --mode argument supports base, fast and base-nightly. Note that base-nightly can be changed without any notice.
• [2023.10.19] Webcam support is not stable currently. We remove the dependency for the latest release. Install with extra dependency option pip install transparent-background[webcam] if you want to use webcam input.
• [2024.02.14] I added a github sponsor badge. Please help maintaining this project if you think this package is useful!
• [2024.08.22] ComfyUI-Inspyrenet-Rembg is implemented by john-mnz. Thank you for sharing great work!
• [2024.09.06] transparent-background total download counts reached 500,000 and ranked 5969 on 🏆top=pypi-package. Thank you all for your huge support!

* [2024.10.05] `--format`, `--resize` and `--reverse` options are implemented.

:inbox_tray: Installation

Dependencies (python packages)

package	version (>=)
pytorch	1.7.1
torchvision	0.8.2
opencv-python	4.6.0.66
timm	0.6.11
tqdm	4.64.1
kornia	0.5.4
gdown	4.5.4
pyvirtualcam (optional)	0.6.0

Note: If you have any problem with pyvirtualcam, please visit their github repository or pypi homepage. Due to the backend workflow for Windows and macOS, we only support Linux for webcam input.

Dependencies

1. Webcam input

We basically follow the virtual camera settings from pyvirtualcam. If you do not choose to install virtual camera, it will visualize real-time output with cv2.imshow.

A. Linux (v4l2loopback)

# Install v4l2loopback for webcam relay
$ git clone https://github.com/umlaeute/v4l2loopback.git && cd v4l2loopback
$ make && sudo make install
$ sudo depmod -a

# Create virtual webcam
$ sudo modprobe v4l2loopback devices=1

• Note: If you have any problem with installing v4l2loopback, please visit their github repository.

B. Windows (OBS)

Install OBS virtual camera from install OBS.

C. macOS (OBS) [not stable]

Follow the steps below.

• Install OBS.
• Start OBS.
• Click "Start Virtual Camera" (bottom right), then "Stop Virtual Camera".
• Close OBS.

2. File explorer for GUI

A. Linux

You need to install zenity to open files and directories on Linux

sudo apt install zenity

Install transparent-background

• Note: please specify extra-index-url as below if you want to use gpu, particularly on Windows.

Install from pypi

pip install --extra-index-url https://download.pytorch.org/whl/cu118 transparent-background # install with official pytorch

With webcam support (not stable)

pip install transparent-background[webcam] # with webcam dependency

Install from Github

pip install --extra-index-url https://download.pytorch.org/whl/cu118 git+https://github.com/plemeri/transparent-background.git

Install from local

git clone https://github.com/plemeri/transparent-background.git
cd transparent-backbround
pip install --extra-index-url https://download.pytorch.org/whl/cu118 .

Install CPU version only

# On Windows
pip install transparent-background
pip uninstall torch torchvision torchaudio
pip install torch torchvision torchaudio

# On Linux
pip install transparent-background
pip uninstall torch torchvision torchaudio
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu

[New] Configuration

transparent-background now supports external configuration rather than hard coded assets (e.g., checkpoint download url).

• The config file will be added in your home directory ~/.transparent-background/config.yaml by default. The directory location can be customized by setting the desired file path under the environment variable TRANSPARENT_BACKGROUND_FILE_PATH. (Contributed by kwokster10)
• You may change the url argument to your Google Drive download link. (Please note that only Google Drive is supported.)
• You may change the md5 argument to your file's md5 checksum. Or, set md5 to NULL to skip verification.
• You may add http_proxy argument to specify the proxy address as you need. If your internet connection is behind a HTTP proxy (e.g. http://192.168.1.80:8080), you can set this argument. (Contributed by bombless)

base:
  url: "https://drive.google.com/file/d/13oBl5MTVcWER3YU4fSxW3ATlVfueFQPY/view?usp=share_link" # google drive url
  md5: "d692e3dd5fa1b9658949d452bebf1cda" # md5 hash (optional)
  ckpt_name: "ckpt_base.pth" # file name
  http_proxy: NULL # specify if needed (Contributed by bombless)
  base_size: [1024, 1024]

fast:
  url: "https://drive.google.com/file/d/1iRX-0MVbUjvAVns5MtVdng6CQlGOIo3m/view?usp=share_link"
  md5: NULL # change md5 to NULL if you want to suppress md5 checksum process
  ckpt_name: "ckpt_fast.pth"
  http_proxy: "http://192.168.1.80:8080"
  base_size: [384, 384]

• If you are an advanced user, maybe you can try making custom mode by training custom model from InSPyReNet.

custom:
  url: [your google drive url]
  md5: NULL
  ckpt_name: "ckpt_custom.pth"
  http_proxy: "http://192.168.1.81:8080"
  base_size: [768, 768]

$ transparent-background --source test.png --mode custom

:pencil2: Usage

:+1: GUI

You can use gui with following command after installation.

transparent-background-gui

:computer: Command Line

# for apple silicon mps backend, use "PYTORCH_ENABLE_MPS_FALLBACK=1" before the command (requires torch >= 1.13)
$ transparent-background --source [SOURCE]
$ transparent-background --source [SOURCE] --dest [DEST] --threshold [THRESHOLD] --type [TYPE] --ckpt [CKPT] --mode [MODE] --resize [RESIZE] --format [FORMAT] (--reverse) (--jit)

• --source [SOURCE]: Specify your data in this argument.
  • Single image - image.png
  • Folder containing images - path/to/img/folder
  • Single video - video.mp4
  • Folder containing videos - path/to/vid/folder
  • Integer for webcam address - 0 (e.g., if your webcam is at /dev/video0.)
• --dest [DEST] (optional): Specify your destination folder. Default location is current directory.
• --threshold [THRESHOLD] (optional): Designate threhsold value from 0.0 to 1.0 for hard prediction. Do not use if you want soft prediction.
• --type [TYPE] (optional): Choose between rgba, map green, blur, overlay, and another image file. Default is rgba.
  • rgba will generate RGBA output regarding saliency score as an alpha map. Note that this will not work for video and webcam input.
  • map will output saliency map only.
  • green will change the background with green screen.
  • white will change the background with white color. -> [2023.05.24] Contributed by carpedm20
  • '[255, 0, 0]' will change the background with color code [255, 0, 0]. Please use with single quotes. -> [2023.05.24] Contributed by carpedm20
  • blur will blur the background.
  • overlay will cover the salient object with translucent green color, and highlight the edges.
  • Another image file (e.g., samples/backgroud.png) will be used as a background, and the object will be overlapped on it.
• --ckpt [CKPT] (optional): Use other checkpoint file. Default is trained with composite dataset and will be automatically downloaded if not available. Please refer to Model Zoo from InSPyReNet for available pre-trained checkpoints.
• --mode [MODE] (optional): Choose from base, base-nightly and fast mode. Use base-nightly for nightly release checkpoint.
• --resize [RESIZE] (optional): Choose between static and dynamic. Dynamic will produce better results in terms of sharper edges but maybe unstable. Default is static
• --format [FORMAT] (optional): Specify output format. If not specified, the output format will be identical to the input format.
• --reverse (optional): Reversing result. In other words, foreground will be removed instead of background. This will make our package's name transparent-foreground! :laughing:
• --jit (optional): Torchscript mode. If specified, it will trace model with pytorch built-in torchscript JIT compiler. May cause delay in initialization, but reduces inference time and gpu memory usage.

:crystal_ball: Python API

• Usage Example

import cv2
import numpy as np

from PIL import Image
from transparent_background import Remover

# Load model
remover = Remover() # default setting
remover = Remover(mode='fast', jit=True, device='cuda:0', ckpt='~/latest.pth') # custom setting
remover = Remover(mode='base-nightly') # nightly release checkpoint

# Usage for image
img = Image.open('samples/aeroplane.jpg').convert('RGB') # read image

out = remover.process(img) # default setting - transparent background
out = remover.process(img, type='rgba') # same as above
out = remover.process(img, type='map') # object map only
out = remover.process(img, type='green') # image matting - green screen
out = remover.process(img, type='white') # change backround with white color
out = remover.process(img, type=[255, 0, 0]) # change background with color code [255, 0, 0]
out = remover.process(img, type='blur') # blur background
out = remover.process(img, type='overlay') # overlay object map onto the image
out = remover.process(img, type='samples/background.jpg') # use another image as a background

out = remover.process(img, threshold=0.5) # use threhold parameter for hard prediction.

out.save('output.png') # save result

# Usage for video
cap = cv2.VideoCapture('samples/b5.mp4') # video reader for input
fps = cap.get(cv2.CAP_PROP_FPS)

writer = None

while cap.isOpened():
    ret, frame = cap.read() # read video

    if ret is False:
        break
        
    frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) 
    img = Image.fromarray(frame).convert('RGB')

    if writer is None:
        writer = cv2.VideoWriter('output.mp4', cv2.VideoWriter_fourcc(*'mp4v'), fps, img.size) # video writer for output

    out = remover.process(img, type='map') # same as image, except for 'rgba' which is not for video.
    writer.write(cv2.cvtColor(np.array(out), cv2.COLOR_BGR2RGB))

cap.release()
writer.release()

:tv: Tutorial

rsreetech shared a tutorial using colab. [Youtube]

:outbox_tray: Uninstall

pip uninstall transparent-background

:page_facing_up: Licence

See LICENCE for more details.

Acknowledgement

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No.2017-0-00897, Development of Object Detection and Recognition for Intelligent Vehicles) and (No.B0101-15-0266, Development of High Performance Visual BigData Discovery Platform for Large-Scale Realtime Data Analysis)