Identification of an Animal Footprint with time prediction using Deep learning
C. Kavitha, C. Hemanath, B. Praveen Raj, N. Sridevi, C. Hemalatha · 2023
This research work addresses issues related to developing an effective and efficient Animal Footprint, classification and tracking system. Successful attempts towards developing algorithmic models for animal segmentation, Animal Footprint, classification and tracking are made. To support the effective classification of animals, two methods are proposed to segment animal image from its background. The proposed animal segmentation algorithm is evaluated using region-based performance measures. A classification model is also proposed based on different features and classifiers. The different features like colour, Gabor and LBP are extracted from the segmented animal images. In addition, a system has been proposed to classify animals in images and videos by using a deep-learning approach. Initially, features are extracted from images/frames using AlexNet pre-trained probabilistic neural network. Later, the extracted features are fed into a multi-class probabilistic neural network classifier for classification.