Design and Implementation - Segmentation of Hand Bone for Bone Age Assessment - page 99

Input hand bone
radiograph

Proposed
MBOBHE

Automated Scale
Selection and

Output: equalized Bone Image

Parameter-free
Diffusion Function

Output: Gaussian Mixture Model

Modified Anisotropic Diffusion

Preprocessing

Output: Diffused Bone Image

Output: Diffused Bone Image

Output: Histogram

Proposed Fuzzy Quadruple Block's Division
Algorithm

Proposed Image
entropy based
Edge detection

Output:texture Information

Proposed Texture- based Crossed Reconstruction

Main Processing Algorithm -Adaptive Crossed Reconstruction

Output: ACR Segmented set of blocks

BARNAE Output: edge's pixels

Quality Assurance Process

Aggregate blocks into full
segmented Image

Fig. 3.29 The overview of the proposed hand bone segmentation framework

has been designed to be automated in terms of the scale selection and the diffu-
sion strength function; this is important to assure that the entire framework has
least human intervention. The diffused and equalized radiograph will then enter
the stage of ACR segmentation using the fuzzy quadruple block's division scheme.
After that, each block undergoes quality assurance process to restore lost image
data and eliminate abundant image data. Finally, all the blocks are aggregated
to form the final segmented hand bone radiograph that are suitable to be used as
input radiograph for any computer-aided skeletal age scoring system. Figure 3.29
summarizes all the abovementioned modules to provide an overview of the entire
segmentation framework.