utkarsh87 / seam-carving

content-aware image resizing

Seam-Carving

Python implementation of Content-Aware Image Resizing.
Uses a dynamic programming approach to find the minimum energy "seam" to "carve" out from the image.

---

Run the script

To view command line options and default arguments:

foo@bar:~/Seam-Carving$ python src/carver.py -h

To run the script on a test image:
Ensure that the test image is in the "images" dir. The results dir will be automatically created if it doesn't already exist and all the results will be saved here. A gif of the entire process will be saved in the current directory as well.

foo@bar:~/Seam-Carving$ python src/carver.py --image_name castle --num_seams 100 --filter_type laplace

---

TODO:

• reduce I/O operations during seam overlaying operation, should ideally reduce compute time
• add gif, current gif too big
• add option to remove horizontal seams
• instead of asking number of seams from user, ask required image dimensions and resize accordingly