image processing: how to separate back ground(black) from the foreground objects (white)

i have tried 2 predefined functions such as im2bw and imbinarize but these functions are resulting in losing part of the objects that exists inside the images. if there is any other way of seperating the objects from the back ground please let me know.

Best regards

Answer 1

[Divya Gaddipati]

Depending on the complexity of the image, different segmentation techniques can be used:

Watershed segmentation: Watershed segmentation technique aims to divide the image into catchment basins and fill them up until they join. MATLAB's implementation uses the most common algorithm for watershed, called Fernand Meyer algorithm

https://www.mathworks.com/help/images/marker-controlled-watershed-segmentation.html

Active Contours: Active Contours (or Snakes) tries to match a deformable model to an image by means of energy minimization. Snakes in particular is designed to solve problems where the approximate shape of the boundary is known. It is a very good technique for motion tracking, when you have identified your model and then you just need to deform the edges a bit to keep tracking the object.

https://www.mathworks.com/help/images/ref/activecontour.html#btuep4x-7

Fast Marching Segmentation: As Snakes, it also requires an initial mask. It models the evolution of a closed curve as a function of time T with speed F(x) in the normal direction at a point x on the curve. The speed function is specified, and the time at which the contour crosses a point x is obtained by solving the equation.

https://www.mathworks.com/help/images/ref/imsegfmm.html

You can also refer to the below page for more techniques:

https://www.mathworks.com/discovery/image-segmentation.html

Hope this helps!