CS240 Seam Carving Project

This project implements dynamic-programming-based seam carving for content-aware image resizing, along with baseline comparisons and report-oriented experiment outputs.

Setup

cd /home/xgy/hw2/cs240_seam_carving_project
python3 -m pip install -r requirements.txt

Run Tests

cd /home/xgy/hw2/cs240_seam_carving_project
pytest -q

Prepare Real Dataset

cd /home/xgy/hw2/cs240_seam_carving_project
python3 scripts/prepare_sample_dataset.py

Dataset source notes:

• See data/BSDS300_DOWNLOAD_MANIFEST.md
• The script downloads the BSDS300 image archive and extracts a fixed 24-image subset into data/input/bsds300_subset/.

Run Single Image

cd /home/xgy/hw2/cs240_seam_carving_project
python3 scripts/run_single.py \
  --image ./data/input/bsds300_subset/159029.jpg \
  --target-width 384 \
  --output-dir ./data/output/manual_single

Run Batch Experiments

cd /home/xgy/hw2/cs240_seam_carving_project
python3 scripts/run_experiments.py

Build Report Assets

cd /home/xgy/hw2/cs240_seam_carving_project
python3 scripts/make_report_assets.py

Output Layout

• data/output/resized/: resized images for each method and ratio
• data/output/plots/: runtime plots
• data/output/tables/: experiment summaries
• reports/figures/: report-ready figures
• reports/tables/: report-ready tables or exported summaries