image-stitcher

This is a python implementation for stitching images by automatically searching for overlap region.

• 👨‍💻 Usage
• 🎯 Preview of results
• 🧠 Main Idea
• 🙋‍♂️ Support
• ✉️ Contact

👨‍💻 Usage

from src.main import main
from src.utils.visualizer import result_visualize

merged_image, cand = main(
    image1=image1,              # The first image to be combined
    image2=image2,              # The second image to be combined
    min_overlap=(5, 5),         # The minimum overlap region
    verbose=False,              # Whether to print the log
)
result_visualize(
    image1=image1,              # The first image to be combined
    image2=image2,              # The second image to be combined
    merged_image=merged_image,  # The output image
    cand=cand,                  # The parameters
)

🎯 Preview of results

The results using CIFAR-10 are shown below. I would refer you to tutorial.ipynb for detailed results.

Figure 1. The example of input images. The red area represents an empty region. This application can combine these images while considering their rotation.

Figure 2. The preprocessed input images. This rotation process is necessary to accurately combine the images. The green frame represents the overlap region between the input images.

Figure 3. The output image.

🧠 Main Idea

Figure 4. The overview of this application in limited case.

This application is designed based on the overlap region's width $w_c$ and height $h_c$. Thanks to this idea, we can simply limit the search space, thus preventing it from capturing overly small, suboptimal overlap region.

Figure 5. The overview of this application.

However, the above approach is not always applicable, specifically when $\min(h_1, h_2) < h_c$ or $\min(w_1, w_2) < w_c$. To address this issue, I change the perspective of $w_c$ and $h_c$ like the above figure. Therefore, this application can handle images of arbitrary sizes.

🙋‍♂️ Support

💙 If you like this app, give it a ⭐ and share it with friends!

✉️ Contact

💥 For questions or issues, feel free to open an issue. I appreciate your feedback and look forward to hearing from you!