Browsing by Author "Wrobleski, Brad"

Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Lucida: Enhancing the creation of photography with an Augmented Intelligence Digital Photography Agent
    (2018-10-22) Wrobleski, Brad; Gadbois, Denis; Jacob, Christian J.; Leblanc, Jean-René
    This thesis explores the research, design and development of a mobile intelligent agent to assist photographers. This work explores the integration of Augmented Reality (AR), Affective Computing (AC), Natural Language Processing (NLP), Computer Vision (CV), and more specifically, Machine Learning (ML) to make possible, the development of a mobile, sympathetic, ambient (virtual), augmented intelligence application (Agent) for assisting photographers. For this research, a prototype agent was designed (Lucida) and developed to assist photographers to enhance the understanding and creation of photographic metering and composition. Learning the art of photography is complicated because of the technical complexity of the camera, the limitations of the user to see photographically, and the lack of real time instruction, feedback and emotive support. A study was completed to measure the effect of how an artificially intelligent agent could communicate photography information. The study explored the communication and pedagogical patterns between human instructor and student and the disparity between human ability and the camera [71]. It analyzed how to transmute the communication and pedagogical values of the human to human experience to the digital milieu using Unsupervised Learning. This research specifically explored how an agent could be trained using the values and attributes of the act of photography and mimic the relationship with an instructor using artificial intelligence. This research focuses on designing and training the agent to find and identify an objective determination of tones in a photography scene (quantitative) and the subjective identification of composition possibilities (intuitive). This thesis illuminates the effectiveness and cogency of Agent based instruction and communication. Further, it demonstrates that a mobile, semantic, sympathetic, augmented, ambient agent can ameliorate the practice of photography in real time, on location by acting as the virtual intelligence to guide photographers. It shows that the integration of specific technologies, and design, produces an effective architecture for the creation of intelligent agent-based assistance.