udrasht/seam-carving-algorithm-for-content-aware-image-resizing

Image resizing using Seam Carving Algorithm

Image_Resizing

Image Resizing using Seam Carving Algorithm

Problem Statement:

Apply seam carving content aware image-resizing algorithm on a given image. Take the height and width (in pixels) of the output image as inputs from the user.

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.
• A vertical seam in an image 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.
• Steps:
  • Energy Calculation: Each pixel has some RGB values. Calculate energy for each pixel. For ex.- You can use dual-gradient energy function but you are free to use any energy function of your choice. Also, you can refer to this link for details.
  • Seam Identification: Identify the lowest energy seam.
  • Seam Removal: Remove the lowest energy seam.

Program Flow:

• Extract individual pixel’s RGB values from the sample image ./data/input/sample.jpeg and write them into ./data/input/rgb_in.txt
• Load the RGB values from ./data/input/rgb_in.txt in a 3D matrix.
• Apply seam carving algorithm.
• Write the individual pixel’s RGB values for resized image into the ./data/output/rgb_out.txt
• Generate sample image output ./data/output/sample_out.jpeg using the ./data/output/rgb_out.txt.

Dependencies:

• You’ll need to install the python image library Pillow to extract RGB values of each pixel and to generate images back from RGB values of each pixel.
• To install Pillow: pip install Pillow

The python script is only compatible with Linux/Mac operating systems.

How to Run::

• python3 driver.py <input_mage_name>
• Ex: python3 driver.py sample1.jpeg

---

Output Example

• Sample Image1

  Original Image

  

  Resized Image

  

• Sample Image2

  Original Image

  

  Resized Image

  

• Sample Image3

  Original Image

  

  Resized Image