Medycyna Wet. 66 (9), 609-613, 2010
| Dejewska T., Boniecki P., Jaśkowski J.M., Kicuła M., Racewicz P. |
| Identification of biometric features of animals through the use of neural image analysis |
| The aim of this study was to examine the possibility of using characteristic features (physical and behavioral) of selected animals in the process of creating neural classification models. A set of specific properties constitutes the basis for the use of a classifier in the form of an artificial neural network. Its construction is based on the need to obtain information encoded in the form of an analyzed image. In this study, examples of potential applications of image neural analysis were presented and possibilities of using this technique in the field of veterinary medicine were proposed. Examples of potential applications were listed. The method of identification and description of characteristic areas visible in the ultrasound image of a cow's uterus were characterized. The need for the skillful gathering of empirical graphic data was considered as vital in order to conduct the extraction of features and the creation of a self-learning representative set. |
| Keywords: artificial intelligence, artificial neural network, neural image analysis, biometrics, image ultrasound |