Objective: To assess possibillity of classification of individuals over and under the age of 30 using CT scans of the medial part of the clavicle by a automatic image analysis. Methods: MSCT scans of the thorax an abdomenfrom KBC Split have been processed through the OsiriX MD 12.program.Afterwards the images were processed through the Gimp 2.10.32. program and trimmed to a size of 700x700 px. A sample of 205 images was divided into a training dataset of 165 images and a model testing dataset of 40 images. In order to make the research as objective as possible a 5-fold cross-validation was applied. The images were loaded into an Orange Data Mining 3.32.0 program for visual programming and data anallysis, which contains several different neural networks and all were tested on the same set of images inorder to select to select the ones that give the most accurate results. Painters gave the best results. Exploratory data analysis was performed using PCA and applying the elbow method.The crossvalidation sample was tested on different classification algorithms and those that give accuracy results higher than 80% were selected. The selected classification methods were tested with a standard threshold value where the posterior probabiloty is 0.95. The results of the anallysis are represented in a binary classification using two classes. Results: Based on the PC1 and PC2 individuals were grouped under 30 and 40 years, between 40 and 55 years and over 80 years. The overall accuracy for model training for all classification algorithms was in the range from 82,5 % to 92,5 %. The most errors in the classification of individuals were observed at the border age ( from 20 to 39 years old). The total accuracy on the test model for posterior probability of 0,5 for SVM was 82,5 %, for LR 92,5 % and for NNI-SGD 90 %, and for posterior probability of 0,95 % the total accuracy of the algorithms LR and NNI-SGD was 100 %, but SVM classified only one person and it did it correctly. Conclusion: The PCA analysis performed on the clavicle scans confirms the possibility of categorizing people according to age. Automatic image analysis achieved classification accuracy between 82,5 % and 92,5 % with a posterior value of 0,5 and maximum accuracy of 100 % with a posterior value of o,95 for all classification algorithms.