High-Fidelity Python Face Swapper

I have two sets of images for you to work with: a fixed-background target photo and a source photo whose face needs to be transplanted. I’m after a fully automatic workflow that detects, aligns, swaps, and blends the face so convincingly that casual viewers won’t notice the edit—even under varied skin tones or lighting. Lighting transfer, colour correction, and seamless edge blending are all essential.

You may combine classical computer-vision routines in OpenCV with deep-learning detectors such as Dlib or MTCNN for best accuracy. The final script should run end-to-end from the command line:

python swap.py --source source.jpg --target target.jpg --output result.jpg

No manual landmarks or masks; everything should be handled programmatically.
the person who with doctor costume will be the main source of images and the other images will be as target image to get it's final swaping .
Deliverables
• Clean, well-commented Python code (3.8+), organised in logical modules
• A requirements.txt (or environment.yml) so the project installs in one step
• The generated result images for the sample pairs I provide
• Brief README explaining model choices, usage, and any optional parameters

Acceptance criteria
• Source face is centred, scaled, and rotated correctly on the target head
• Skin tone and lighting match the target scene without visible artefacts
• Background remains untouched; no ghosting or colour spill
• Script completes in a reasonable time on a standard GPU-free machine

If this aligns with your skill set, I look forward to reviewing your approach and time estimate once you’ve examined the attached doc and sample images.