Academia.eduAcademia.edu

Figure 1 – uploaded by International Journal of Computer Science & Information Technology (IJCSIT)

See full PDFdownloadDownload figure
Table 3.4 Skin burn characteristics Fig 3.3 Full thickness burn

Table 3 4 Skin burn characteristics Fig 3.3 Full thickness burn

Related Figures (11)

Fig 3.1 Superficial dermal burn Superficial burn means injury to the epidermal layer of the skin which normally heals within 5 to 7 days. A common type of superficial burn is sunburn. Figure 3.1 is a example of superficial burn.
Fig 3.1 Superficial dermal burn Superficial burn means injury to the epidermal layer of the skin which normally heals within 5 to 7 days. A common type of superficial burn is sunburn. Figure 3.1 is a example of superficial burn.
Fig 3.2 Partial thickness burn In the partial thickness burn the damage extends up to the dermis layer. As a result, the seriousness of a partial thickness burn depends on how much the dermis has been injured, these burns usually leave scars. Figure 3.2 is an example of partial thickness burns.
Fig 3.2 Partial thickness burn In the partial thickness burn the damage extends up to the dermis layer. As a result, the seriousness of a partial thickness burn depends on how much the dermis has been injured, these burns usually leave scars. Figure 3.2 is an example of partial thickness burns.
In this method of classification comparison of data values is done. Templates belonging to different class of burn images are compared with the new samples. The sample image is classified into the class, whose template matches with the new sample with least error. Figure 4.1 illustrates the template matching method in which the features under selected region are compared with the previously stored values. The process is inherently such that it takes more time to process compute and throughout results when the sample size is more [5]. Figure 4.1: Template Matching Method
In this method of classification comparison of data values is done. Templates belonging to different class of burn images are compared with the new samples. The sample image is classified into the class, whose template matches with the new sample with least error. Figure 4.1 illustrates the template matching method in which the features under selected region are compared with the previously stored values. The process is inherently such that it takes more time to process compute and throughout results when the sample size is more [5]. Figure 4.1: Template Matching Method
Figure 4.2: Choosing k value Given a training set D and a test object z, which is a vector of attribute values and has an unknown class label, the algorithm computes the Euclidean distance between z and all the training objects to determine its nearest neighbor list. It then assigns a class to z by taking the class of the majority of neighboring objects. Ties are broken in an unspecified manner, for example, randomly or by taking the most frequent class in the training set.
Figure 4.2: Choosing k value Given a training set D and a test object z, which is a vector of attribute values and has an unknown class label, the algorithm computes the Euclidean distance between z and all the training objects to determine its nearest neighbor list. It then assigns a class to z by taking the class of the majority of neighboring objects. Ties are broken in an unspecified manner, for example, randomly or by taking the most frequent class in the training set.
International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 1, February 2013 The data for training is a set of points (vectors) x;,, along with their categories y;. For some dimension d, the n x;eR“ , and the y; = ¥1. The equation of a hyperplane is < w,x > +b = 0 where, weR“, < w,x > is the inner dot product of w and b is real.
International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 1, February 2013 The data for training is a set of points (vectors) x;,, along with their categories y;. For some dimension d, the n x;eR“ , and the y; = ¥1. The equation of a hyperplane is < w,x > +b = 0 where, weR“, < w,x > is the inner dot product of w and b is real.
B. IMAGE CLASSIFICATION
B. IMAGE CLASSIFICATION
Table 5.2: Template Matching Method
Table 5.2: Template Matching Method
Figure 6.1 shows the chrominance plot of features, mean and DCT values of 120 skin bum images belonging to 3 different categories. Features of 40 images belonging to each category are plotted. (2,1)"" coefficient of DCT values are plotted on x-axis and chrominance mean on y-axis. The distribution of these features allows the classifiers to learn and discriminate the test sample values into different category. In this method of classification, a new test vector containing mean and DCT values of sample images is matched with the previously stored templates belonging to different grades. Depending on the closeness of the match, test sample is categorized. Testing of 120 images, with 40 images from each category is carried out. 20 images from each category are used for training and remaining 20 for testing with cross validation. Overall a classifier efficiency of 66% is incurred with this method, the results are tabulated as shown in the Table 5.2.
Figure 6.1 shows the chrominance plot of features, mean and DCT values of 120 skin bum images belonging to 3 different categories. Features of 40 images belonging to each category are plotted. (2,1)"" coefficient of DCT values are plotted on x-axis and chrominance mean on y-axis. The distribution of these features allows the classifiers to learn and discriminate the test sample values into different category. In this method of classification, a new test vector containing mean and DCT values of sample images is matched with the previously stored templates belonging to different grades. Depending on the closeness of the match, test sample is categorized. Testing of 120 images, with 40 images from each category is carried out. 20 images from each category are used for training and remaining 20 for testing with cross validation. Overall a classifier efficiency of 66% is incurred with this method, the results are tabulated as shown in the Table 5.2.
Figure 5.1: Classification of gradel and grade2 vectors International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 1, February 2013
Figure 5.1: Classification of gradel and grade2 vectors International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 1, February 2013
Figure 5.2: Classification of gradel and grade3 vectors Figure 5.3: Classification of grade2 and grade3 vectors
Figure 5.2: Classification of gradel and grade3 vectors Figure 5.3: Classification of grade2 and grade3 vectors
Related topics:
Computer science & Information TechnologyComputer Science and Information TechnologyInformation Technology/computer ScienceInformation Technology and Computer ScienceComputer Science Information Technology
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved