The behavior of animals has been commonly recorded in a manual way. Traditionally, if the researcher considers that a certain behavior is displayed, he or she notes the behavior, either by writing it down or by entering the data into an event-recording program. Automated observation with the aid of video cameras and computer programs provides significant advantages, it is more reliable and continuous. Video tracking is particularly suitable for measuring locomotor behavior, expressed as spatial measurements (distance, speed, turning, etc.) that the human observer is unable to accurately estimate. Automated observation systems also allow the study of behaviors that occur briefly and are then interspersed with long periods of inaction. The recording, analysis and classification of animal behavior has evolved dramatically in the past decade. Video tracking is a powerful new tool for studying natural undisturbed behavior in a wide range of animals. It employs a hardware setup that integrates (several) video cameras (and microphones) with computer vision tools to track the movements of animals and to classify well-described behaviors. Such accurate, quantitative measurement of animal behavior is critical, not only for researchers in academic institutions studying animal behavior, but also for veterinary medicine (and pharmaceutical companies) developing drugs to treat disorders affecting social interactions. Our system has been developed with an Austrian company (LoopBio) and consists of several multi-camera tracking systems (for manakins, ravens, dogs, pigs), with automatic identification of unmarked animals, and multi-audio system for recording animal vocalizations. The system features video management, annotation (incl. synchronized annotations for 3D reconstruction with multi camera recording systems), behavioral scoring/coding, image processing, modules for detecting and tracking objects, flowtrace rendering, Deep Learning and AI modules to detect single and multiple individuals, keypoint / pose tracking (to detect multiple key points on a single animal), and finally integration into local network storages.
Prof. Ludwig Huber
If you are interested to use the infrastructure (e.g. scientific project or collaboration), please contact us. For possible cooperation contact Ludwig Huber (Ludwig.email@example.com)
Methoden & Expertise zur Forschungsinfrastruktur
Behavioral and cognitive biologists study the behavior of animals in captive settings as well as in the wild. They need to detect, analyse and interpret animal behavior in order to investigate the underlying mechanisms and procedures. Automated observation with the aid of video cameras, microphones, computers and machine learning algorithms provide significant advantages. Behaviors are recorded more precisely (in space and time), more reliably and for a long time (the system does not suffer from observer fatigue or drift). The involved researchers at the University of Veterinary Medicine and the University of Vienna study the behavior and cognitive abilities of a broad range of species (birds and mammals) in a great variety of environments (lab, aviary, bit enclosures, in the wild). They have developed the new systems together with an Austrian company (LoopBio), have installed them at different research locations (in Vienna and Lower Austria) and will continue to create new applications and programs.