Beyond Show of Hands: Engaging Viewers via Expressive and Scalable Visual Communication in Live Streaming

Title of the Paper

Beyond Show of Hands: Engaging Viewers via Expressive and Scalable Visual Communication in Live Streaming

Paper Information

• Subject Area: Live streaming interaction and visual communication
• Keywords: live streaming interaction, user engagement, drawing, visualization, visualization tools, user experience, live streaming software, real-time visualization, viewer interaction, data aggregation

Research Background and Issues

• Identified Problems/Challenges:

  1. Existing live streaming interactions are primarily text-based, lacking the ability to express visual and spatial information.
  2. Visual interaction systems face two major challenges: interpreting large-scale viewer inputs and maintaining expressive and adaptive interaction.
  3. Structuring viewers' visual inputs may limit interaction diversity, while overly flexible inputs may lead to information chaos.

• Research Importance: With the growing prevalence of online live streaming and video conferencing in fields such as education, gaming, and artistic creation, developing an interaction method that ensures scalable viewer participation while balancing expressiveness and adaptability can significantly enhance engagement and interaction quality.

• Research Motivation and Related Work:

  1. Existing studies, such as text aggregation, simple clicks, and tools for viewers to select content in live streaming scenarios, fail to simultaneously achieve viewer expressiveness and adaptability to live content.
  2. Utilizing visual input tools can significantly enrich live streaming interactions, showing potential in enhancing viewer engagement and comprehension.

Solution

• Proposed Solution: Introducing the Visual Input Management Framework (VIMF) and a specific system—VisPoll. VIMF aims to assist hosts in specifying, aggregating, and visualizing viewers' visual inputs during live streaming, as well as interpreting the information.

• Innovations:

  1. Structuring viewer inputs using "visual attributes," including shape, color, size, rotation, and position, to flexibly control and manage visualization.
  2. Allowing viewers flexibility in expressing visual information while simplifying management for hosts.
  3. Supporting aggregation of viewer inputs and associating them with live content for display.

• Implementation Steps and Techniques:

  1. Input Specification: Hosts define areas viewers can interact with, visual symbols (e.g., oxygen or hydrogen images), and restrictions on visual attributes.
  2. Input Aggregation: Visual data is processed using clustering algorithms (e.g., hierarchical clustering) to classify viewer inputs based on specific attributes.
  3. Data Visualization: Aggregated results are visualized through axial layouts and glyphs, enabling viewers to intuitively understand others' inputs.

Research Outcomes

• Specific Outcomes:

  1. The initial VisPoll system successfully achieved scalable and expressive viewer interaction in live streaming.
  2. Integrated viewer input aggregation and real-time visualization features improved interaction efficiency in live streaming.
  3. Collaboration with educators, creators, and gaming streamers validated the system's applicability across multiple domains.

• Comparison with Existing Solutions:

  1. The system is compatible with diverse live streaming content, not limited to specific environments.
  2. It provides viewers with richer visual expression methods beyond simple clicks or text inputs.
  3. Automated aggregation and visualization significantly reduce the burden on hosts while enhancing viewer engagement.

• Experiment and Evaluation Results: Preliminary experiments in three domains (creative live streaming, educational lectures, gaming live streaming) demonstrated that VisPoll fosters more active viewer participation while helping hosts quickly interpret collective feedback. Both qualitative and quantitative data indicate improved viewer engagement and experience.

• Limitations and Future Directions:

  1. The system requires hosts to perform complex configurations in advance, which may demand high technical proficiency.
  2. Clustering algorithms occasionally fail to classify viewer inputs as expected.
  3. Currently, the system only supports static visual attributes and has not extended to dynamic inputs or complex visual object relationships.
  4. Future plans: Deploy more flexible preset templates, reduce the learning curve for hosts, support multi-round visual input collection, and enable dynamic input aggregation and visualization.

Through this research, the authors envision the future of live streaming interaction, emphasizing the importance of scalable and expressive visual communication tools, which may inspire new directions in the field of human-computer interaction.