This study explores the use of deep learning neural networks and transfer learning to estimate the age of sheep from their dental images. This is an important aspect of agriculture for meat quality, animal welfare, breeding, and health management. Using cutting-edge techniques, MobileNet, ResNet50, and ResNet102, we compare two deep learning approaches: fine-tuning and feature extraction using the pre-trained version of these models as part of our investigation. We collected 540 images of sheep from nearby farms, concentrating on three age groups: young, middle-aged, and elderly, for the purpose of our study. With an interesting recognition accuracy of 96.9%, the experimental results demonstrate that ResNet102 is the best performer both when fine-tuned and when employing its deep features that are retrieved from its pre-trained version. These findings highlight how cutting-edge machine learning techniques have the potential to completely transform long-standing methods in the sheep sector and pave the way for developing a novel mobile application that improves economic outcomes and cultural conformity concerning sheep age recognition.

Collaborative Virtual Environments (CVE) with co-located or remote video communication functionality require a continuous experience of social presence. If, at any stage during the experience the communication interrupts presence, then the CVE experience as a whole is affected - spatial presence is then decoupled from social presence. We present a solution to this problem by introducing the concept of a virtualized version of Google Glass™ called Virtual Glass. Virtual Glass is integrated into the CVE as a real-world metaphor for a communication device, one particularly suited for collaborative instructor-performer systems. In a study with 65 participants we demonstrated that the concept of Virtual Glass is effective, that it supports a high level of social presence and that the social presence is rated higher than a standard picture-in-picture videoconferencing approach for certain tasks.