Animal Detection

Automatic Screening of Wildlife Trail-Camera Imagery

The Applied Vision Lab aims to determine the presence of animals in a set of images from the Texas Parks and Wildlife Database (TPWD). The team is currently using the Xception Convolutional Neural Network (CNN) architecture and transfer learning to perform animal detection and (bird) localization.

A gallery of various day and night images tackled by AVL (click to enlarge)

Examples of day and night images

General project workflow and approach

Labelling tool developed by AVL to annotate bird

Automatic localization of birds using a deep learner 

 

Sponsor/Partner : Texas Parks and Wildlife