Bone Fracture Detection Using Morphological and Comparing the Accuracy with Genetic Algorithm Y. Rakesh

Discussion
This experiment was done with 39 and 236 training images separately. For 39 training images, the opening 
values are decreased by 20 % on morphological and morphology does not have a great impact on success rates, 
but 50 % increase the success rates effectively. Trained sample values taken for statistical analysis using IBM 
SPSS software. The P value is 0.029 obtained from the SPSS software tool proves that morphology has better 
accuracy than Genetic in the image recognition system. The comparison of mean accuracy values for two groups 
morphological and Genetic with p-value 0.05 and error bar 95 % with the effective prediction is shown in Fig. 2. 
The error bars with the mean accuracy detection +/- 1 SD. 
A method for moderately long-bone shaft splitting has been given, as well as fracture diagnosis within the 
segmented region.With an accuracy of 91.16 %, effectively diagnosed and accurately indicated the sites for tiny, 
difficult bone fractures in infants responsible for over half and radius bones using local entropy (Hržić et al. 
2019). Employing a combined approach of the gradient that incorporates magnitude and direction data, with line 
parameters derived using a modified Hough transform, to correctly identify fractures within the diaphyseal 
portion of a long bone with 87.26 % accuracy (Jagtap and Holambe 2018). In the test data set, 83 % of the 
diaphysis fragmentation borders and 83 % of the fractures inside those features extracted were successfully 
recognized and detected (Eccles et al. 2020). It is recommended that they complement each other; this may be 
accomplished by completing a significance and consistency analysis of classifiers, where the best features from 
the high-dimensional feature collection can be chosen with 82.23 accuracy (Brownlee 2018). The use of deep 
learning to detect morphological algorithms is viable and can be done with 79.86 % accuracy (Yu et al. 2020). 
Overall diagnostic accuracy of rib cage reformats for the diagnosis of rib fractures was comparable to those of 
traditional reformats, with high repeatability and a significant decrease in assessment time 73.46 % accuracy 
(Urbaneja et al. 2019). 
The major limitation of this proposed method is to enhance the image segmentation under different lighting 
conditions in deep learning. The future scope of this proposed research will be used to improvise the 
performance rate under different conditions with Novel Feature Extraction.