Aim/Purpose
Computer vision is advancing rapidly, but conventional instruction can struggle to match contemporary students’ preferences for visual, interactive, and immersive learning experiences; this work addresses how to design a computer vision course that keeps students engaged in gamified projects while still improving measurable learning outcomes.

Background
This Teaching Practice Research Project (supported by Ministry of Education in Taiwan) designed and evaluated a Gamified Project-Based Learning (GPBL) approach in an undergraduate “Computer Vision” course.

Methodology
A quasi-experimental, nonequivalent groups pretest–posttest design was used: the experimental group received GPBL (game elements embedded into project learning), while the control group received traditional PBL; learning effectiveness and learning motivation were measured as dependent variables, with teaching content and instructor held constant and pretest used for statistical control.

Contribution
This study contributes an implementable GPBL course design for computer vision (modules, sequencing, and classroom orchestration) and provides empirical evidence on its effects in a higher-education, practice-heavy CS context. It also surfaces design tradeoffs between authenticity and constraints (e.g., privacy-safe avatars vs. detection reliability) that matter when “game assets” become learning data.

Findings
Results show a statistically significant improvement in learning effectiveness for GPBL (mean 79.6) over PBL (mean 72.1; p = 0.046), while motivation and satisfaction increased slightly but did not reach statistical significance.

Recommendations for Practitioners
Treat game elements as supporting mechanisms to spark interest and participation, not as the sole driver of learning; keep the focus on authentic “problem solving” value in computer vision tasks.

Recommendations for Researchers
Study which types of game mechanics (feedback loops, progress systems, challenge design) best transfer to technical skills and knowledge retention in computer vision, beyond general “gamification vs. non-gamification” comparisons.

Impact on Society
By improving learning effectiveness in a computer-vision course through a structured GPBL approach, this work supports building a workforce capable of applying fast-evolving vision technologies in real applications, while highlighting responsible design considerations that shape how students learn with data-driven tools.

Future Research
Future work should iterate GPBL modules to better align “game data” with real-world detection conditions (e.g., replacing or augmenting Mii-based tasks with privacy-preserving real videos/open datasets) and experimentally test added scaffolding strategies (hints, exemplar pipelines, staged milestones) to see whether motivation and satisfaction can be improved without reducing challenge.