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

Statistical Analysis 
The software used for this analysis was IBM statistical tools (SPSS) (Jaafar et al. 2020). The Independent-
sample t-test is done with the accuracy values of both algorithms. SPSS software is used for statistical 
analysis(Watkins 2021). Height and Width are independent variables, while accuracy and specificity are 
dependent variables. The performance of algorithms is compared using an independent t-test. 
 
Result 
The input images which are holding for testing and training images with low resolution and noise. Previously 
the image is read and it undergoes preprocessing technique in which the RGB image is converted into grayscale 
to reduce the noise (Ali, Abood, and Khudhair 2019). The accuracy differences of these training and testing 

---

Journal of Pharmaceutical Negative Results ¦Volume13¦SpecialIssue 4¦2022
272
 
Y. Rakesh
,
etal
.:
Bone Fracture Detection Using Morphological and Comparing the Accuracy with Genetic Algorithm 
images between the morphological and genetic algorithm. Morphological gains the accuracy rate of 87.46 % 
compared to the existing genetic technique which achieves 83.25 % accuracy rate. The recognition rates of 
morphological and genetic techniques with different training images on edge detection are represented in Table 
1. Table-2 represents the recognition rates of morphological and morphology with different training images to 
get the values of mean, standard deviation and standard error. The gray variations of test and trained bone 
images are experimented to evaluate the robustness of morphological and Genetic due to the edge detection on 
temporary illumination variations in fracture detection. The results of these techniques are presented in Table 3. 
The bone image is recognized with the help of bounding box and this image is cropped to increase the accuracy 
with a resolution of 256*256 (24-bit jpg) bone image. The analyzed bone image and the cropped image based on 
edge detection is as shown in Fig. 1. Figure 2 shows the accuracy differences of the training dataset between 
morphological and genetic techniques in deep learning. This shows that morphology achieves a higher accuracy 
rate of 87.46 % compared to the existing Genetic technique (Marinov, Kalmukov, and Valova 2021).