Artificial-Eyeliner

A script to artificially apply eyeliner on live video (Webcam videos and images also supported)

Demo:

Before	After

---

Project Structure:

Flow Chart:

Project Explaination:

• We first locate the faces present in an image.
• On each face detected, we locate 68 facial landmark points.
• We extract the points representing eyes (pt no. 37-49) out of those 68 points.
• We interpolate over the extracted eye landmark points to generate a curve.
• We draw over this curve and we have our eyeliner.

Code Overview:

• lndMrkDetector() : To extract 68 facial landmark points
• getEyeLandmarkPts() : To extract eye landmark points (pt no 37-48)
• getEyelinerPoints() : To interpolate those eye landmark points
• drawEyeliner(): To draw eyeliner on interpolated points

---

Getting Started:

Clone the repo using:

git clone https://github.com/kaushil24/Artificial-Eyeliner/

Starting virtual environment:

• Install the venv module:
  python3 -m pip install --user virtualenv
• Setup the virtual environment and activate it:
  python3 -m venv env
source env/bin/activate

Installing the dependencies:

• To install Dlib please visit this article-https://www.pyimagesearch.com/2017/03/27/how-to-install-dlib/
• Rest of the dependencies: pip3 install -r requirements.txt

Sample Test:

python3 eyeliner.py -v "Media/Sample Video.mp4" -s "Output"

---

CLI Usage

python eyeliner.py [-i image] [-v video] [-d dat] [-t thickness] [-c color] [-s save]

• -i : Location of image you want to apply eyeliner on
• -v : Location of video you want to apply eyeliner on.
• -v : Live eyeliner of webcam video if webcam is given (Eg: python3 -v webcam -s "webcam")
• -t : Whole interger number to set thickness of eyeliner. Default = 2. Recommended number value between 0-5
• -d: Path to your shape_predictor_68_face_landmarks.dat file. Default value is the root unless you have the shape_predictor_68_face_landmarks.dat file stored at some other location you need not use this argument.
• -c : Change color of the eyeliner. Use -c 255 255 255. Defaule = 0 0 0.
• -s : Location and file name you want to save the output to. NOTE The program automatically adds extension while saving the file. NOTE: If a file with same name already exists, it will overwrite that file.

---

To Do:

• Analyze the results on different eye shapes and color.
• Dynamically change eyeliner thickness based on the distance between subject and camera.
• Add support for people wearing spectacles.

---

Pull requests are always welcomed :D

Version Info:

• Python - 3.6
• Numpy - 1.17.4
• Dlib - 19.18.0
• cv2 - 4.1.2
• matplotlib - 3.1.2
• skimage - 0.16.2
• scipy - 1.3.3
• imutils - 0.5.3
• PIL - 6.2.1

Credits:

Face images ("Sample Image.jpg") generated by AI. https://generated.photos