Seam-Carving-Algorithm-for-Content-Aware-Image-Resizing

Description

Seam carving is a content-aware image resizing technique where the image is reduced in size by one pixel of height (or width) at a time. It works by identifying and removing the lowest energy seam from the image.

What is Seam Carving?

Seam carving is a content-aware image resizing technique where the image is reduced in size by one pixel of height (or width) at a time. It operates by finding the lowest energy seam in the image and removing it.

• A vertical seam is a path of pixels connected from the top to the bottom with one pixel in each row.
• A horizontal seam is a path of pixels connected from the left to the right with one pixel in each column.
• The algorithm consists of three main steps:
  1. Energy Calculation: Calculate the energy for each pixel using an energy function.
  2. Seam Identification: Identify the lowest energy seam in the image.
  3. Seam Removal: Remove the lowest energy seam from the image.

Output

The generated output images for the sample inputs can be found at ./data/output/sample_out.jpeg. The RGB values used for generating the output images are stored in ./data/output/rgb_out.txt.

Here are some visual comparisons between the input and output images:

Input Image	Output Image

Program Flow

1. Extract individual pixel RGB values from the sample image ./data/input/sample.jpeg and save them in ./data/input/rgb_in.txt.
2. Load the RGB values from ./data/input/rgb_in.txt into a 3D matrix.
3. Implement the seam carving algorithm (inside the solve() function in main.cpp).
4. Save the individual pixel RGB values for the resized image in ./data/output/rgb_out.txt.
5. Generate the output sample image ./data/output/sample_out.jpeg using the RGB values in ./data/output/rgb_out.txt.

For convenience, a Python script ./src/driver.py is provided, which performs the first and fifth tasks. The corresponding ./src/main.cpp file is executed by the Python script to handle the second and fourth tasks.

File structure

• src/
  • main.cpp
  • driver.py
• data/
  • input/
    • sample1.jpeg -output/
    • output.jpeg
• artifacts/
• bin/

Dependencies:

You'll need to install the Python image library Pillow to extract RGB values of each pixel and to generate images from RGB values.

To install Pillow, run the following command:

pip install Pillow

Please note that the provided Python script is only compatible with Linux/Mac operating systems. If you're a Windows user, you can request the input text file (rgb_in.txt) from us to work on the problem.

How to Run:

1. Open the directory in the terminal.
2. Run the Python script driver.py located in /src. Pass the input image filename (must be present inside /data/input) as a command line argument.

Ex:

python ./src/driver.py sample1.jpeg

Note: Make sure to have the necessary input image file (sample1.jpeg in the above example) in the /data/input directory before running the script.

Note: Make sure to replace the example image file name (sample1.jpeg) with the actual image file you want to use.

Reference

The Calculation ofGradient Matrix and details of Seam Carving Algorithm was reffered from an Assignment from Princton University which can be found here

Contributions

Contributions are welcome! If you encounter any issues or have suggestions for improvements, please open an issue or submit a pull request.