Image processing and edge detection on grayscale images

Overall Requirement
To detect the edge of vertebras and leg bone in a grayscale full body image in front view and side view.

Input/Output
1. LF.tif : Low contrast Front view image. Used to find range of spine
2. HF.tif: High contrast Front view image. Used to find details of vertebra and leg bones contour
3. LS.tif: Low contrast Side view image. Used to find range of spine
4. HF.tif: High contrast Side view image. Used to find details of vertebra and leg bones contour

Procedures
- Detect the outline of the spine using low contrast Image
- Apply the ROI to High contrast Image to reduce searching area for spine
- Find out the vertebras. Also find out the leg bone contour.
- Do both front view and side view
 
Final Output


Extra notes
The method should work best-effort with different conditions
- Spine may be straight or curved
- - Ideally, Vertebras are detected in its full form.
- - In worst effort, Vertebra should be presented as 4-line quadrilaterals.
- - Data is precious so training is impossible. Only classical image processing techniques can be used.
- Programming Language can be of Matlab, Python or C, C#

Input files are about 50MB each.