We consider an image with n colored regions. We want to extract these n region to put in n different images to treat them and to build the original image with the same initial position for each region but the content of each image changing.
That will be interesting in infography
Assuming that the individual regions do not overlap:
1) Segment the desired 'n' regions and create 'n' binary masks one for each region.
2) Create 'n' empty bitmaps and preserve the original image size and parameters.
3) Use the binary masks to copy each region into one of the new bitmaps.
4) Apply your individual processing to each of the images.
5) At last use the binary masks to copy all the 'n' bitmaps into one image again.
Not 100% sure if I understood your question correctly but I hope that helps.
Assuming that the individual regions do not overlap:
1) Segment the desired 'n' regions and create 'n' binary masks one for each region.
2) Create 'n' empty bitmaps and preserve the original image size and parameters.
3) Use the binary masks to copy each region into one of the new bitmaps.
4) Apply your individual processing to each of the images.
5) At last use the binary masks to copy all the 'n' bitmaps into one image again.
Not 100% sure if I understood your question correctly but I hope that helps.
The method change with the shape of region, which region are you looking?. A powerfull segmentation algorithm is the one based on active contourns
Binarize your input image.
Label binary image (use bwconncomp if you use matlab)
Create n number of image and assign one connected component to each of these images. Now you can modify each of these images independently and combine them
modify Julius methods a little
1. You can convert rgb image to hsv image, and use n color thresholds to segment n regions (also label the region with its color)
2. Assign one connected component to each of these images
3. modify each of these images independently and combine them
Use floodfill algorithm
1. Discover top and leftmost pixel, and save color (X)
2 Use floodfill with a color based on the (flood while = X; replace X with a unique color)
3 Note the envelope has been in flood areas
4 To 1
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing