Car Segmentation Package

Overview

The CarSegmentPro package is developed to facilitate the removal of both internal and external backgrounds from car images. This package comprises two distinct models:

External Model:

Purpose: Removes the background external to cars.

Performance: Known for its effectiveness in this task.

Internal Model:

Purpose: Removes the background internal to cars.

Note: This task involves innovation, where a dataset was collected, and a deep learning model was trained. The achieved accuracy is 75%, which is considered moderate. Keep in mind that sentiment analysis provided suboptimal results.

Usage

External Model Usage

To use the external model for removing the background external to cars, employ the following code:

from carbgremover.external_model import remove_background_external, plot_image
# Parameters:
# image_path: path to the image
# device: "cpu" (default) or "cuda" if you have GPU
res = remove_background_external(image_path='car1.jpg', device="cpu")
plot_image(res, figsize=(15, 15))

Input image:

Output image: For saving the image, use the following:

import cv2
cv2.imwrite('rescar2.jpg', res)

Internal Model Usage

For the internal model designed to remove the background internal to cars, utilize the following code

from carbgremover.internal_model import remove_background_internal,plot_image

res = remove_background_internal('car2.jpg')
plot_image(res, figsize=(15, 15))

Input image:

Output image:

For saving the image, use the following:

import cv2

res = cv2.cvtColor(res, cv2.COLOR_RGB2BGR)
cv2.imwrite('rescar2.jpg', res)