Object Segmentation using PixelLib

CarScan is a project which focus on segemnetation of object (basically car) in a real-time scenerio. This project uses Semantic Segmentation with help of PixelLib module.

This project uses the pretrained model which can be downloaded from here

Packages Used

1. OpenCV
2. tensorflow
3. numpy
4. pillow
5. pixellib

Running the code

1. Run MainCode.py and the final output image would be dumped into "Output images" directory. The "Input images" directory contains the input image for the program.

Referenece

1. Image Segmentation With 5 Lines 0f Code (https://towardsdatascience.com/image-segmentation-with-six-lines-0f-code-acb870a462e8 )
2. Interactive Foreground Extraction using GrabCut Algorithm (https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_grabcut/py_grabcut.html#grabcut)
3. PixelLib (https://github.com/ayoolaolafenwa/PixelLib)