Neural Networks for Pattern Recognition
Neural Networks for Pattern Recognition
Image Processing, Analysis, and Machine Vision
Image Processing, Analysis, and Machine Vision
| Hi-index | 0.00 |
Jaguar (P. onca) tracks can help in different studies on big felid research. However, tracks can be confused with those of pumas (P. concolor) making it difficult to determine priority areas. There has been some attempts to obtain a formula to differentiate between the two species however this can be time consuming and requires some experience. We propose the use of a computer based image processing system in order to facilitate studies with these species. A set of 44 footprints from jaguar (23) and puma (21) was obtained in Sonora. These footprints were used to generate a program based in a Generalized Linear Model (GLM) that uses two shape descriptors (a boundary based orientability measure and boundary length) to classify the footprints. From the existing data set, the GLM was able to classify the footprints with an accuracy of 75%. A larger data set is expected to increase the classification accuracy. Additional shape descriptors may be added for further improvement. This will provide a method of easy access to carry further studies at a really low cost.