In-field evaluation and behavioral analysis of a drawing-based penguin observation system

In-field evaluation and behavioral analysis of a drawing-based penguin observation system Yuki Nakagawa, Meiji University Satoshi Nakamura, Meiji University

Background Leisure Zoos & Education Aquariums Research Conservation Zoos and aquariums provide not only leisure, but also opportunities for learning and conservation. 2

Background Zoos and aquariums as entertainment contexts Living Animals × × Attention × Interpretation Social Interaction Experiential Entertainment 3

Background • Previous work suggests that identifying individual animals increases empathy and conservation awareness. [Smith+ 2024] • Visitors tend to observe the group as a whole, with limited attention to individual characteristics and behaviors. Smith, P., Mann, J. and Marsh, A.: Empathy for wildlife: The importance of the individual, Ambio, Vol. 53, No. 9, pp. 1269–1280 (2024). 4

Background • Most zoos and aquariums identified individuals using flipper bands. Name-tagged bands Color-coded bands 5

Background • Some zoos and aquariums displayed information about individual. Individual penguin list Penguin relationship diagram (SEA LIFE Sydney) （Kakegawa Kachouen） 6

Method Drawing an abdominal pattern. × Retrieve the name from the drawing. 7

System Design Draw the penguin’s spot pattern. 8

System Design The system retrieves and ranks candidate penguins by similarity. 9

System Design Tap a penguin to view its profile information. 10

System Design Can’t see the right side! 11

System Design Mark invisible areas Can’t see the right side! 12

Catalog Album 14

Previous Studies Controlled Study [AVI 2024] Field Study [CHI 2026] Drawing-based observation: In a aquarium, the system encouraged: • Increased interest in individual penguins • Individual-focused observation • Influenced recognition & memory of individual characteristics • Longer dwell time at the exhibit • Communication among companions Previous evaluations relied mainly on behavioral observations, making it difficult to understand how visitors' drawing-based interactions actually unfolded in the field. Nakagawa & Nakamura, AVI 2024 — Drawing-type Search Method Focusing on Penguin’s Abdominal Patterns Nakagawa & Nakamura, CHI 2026 — Drawing Attention: A Field Study of Sketch-Based Penguin Identification 15

Research Questions How do visitors engage with drawing-based individual identification RQ1: as a playful observation activity in an aquarium setting? How does drawing-based interaction support comparative and RQ2: socially shared observation of individual penguins? RQ3: What behavioral indicators suggest that the experience becomes revisitable within and beyond the immediate exhibit encounter? 16

Field Study • We conducted a field study at Sunshine Aquarium (Tokyo). • 7days: Oct 9–15, 2025 • 47 African penguins • Logs + behavioral observations 17

Findings — Usage and Drawing Behavior 340 8.8s 5.24 browser-level users median drawing time strokes /drawing Visitors could engage with the system through lightweight sketching. 18

Findings — Candidate Comparison Rank of the selected candidate (n = 76 identifications) 25 32.9% chose the top-1 88 44 4 1 1 6 22 1 11 1112 11 16 112 1 21 26 31 1 36 1 1 2 53.9% chose the top-3 41 Visitors did not always simply accept the top result, but explored and compared multiple candidates. 19

Findings — Shared and Repeated Use Repeated Access • 122 users (35.9%) accessed the system multiple times within the same day. • 23 users (6.8%) accessed the system across multiple days. Collaborative Use • Visitors often shared one smartphone. • Family members drew and checked results together. • Companions took turns drawing or used photos taken by others as references. 20

Findings — Revisiting Individual Information • 78.3% of multi-day users accessed the catalog, compared with 28.1% of other users. • Multi-day users also viewed the catalog significantly more often. Repeated use was associated with returning to individual penguin information. 21

Discussion — From Identification to Observational Play Low-effort engagement - Visitors could start by drawing only the features they could see. Exploratory comparison - Multiple ranked candidates encouraged visitors to keep looking and comparing. Socially shared observation - Drawing made observed features visible and discussable among companions. 23

Discussion — From Identification to Observational Play Not longer sessions, but repeated returns - Visitors returned through short, repeated access over time. Returning to individual information - Multi-day users were much more likely to revisit the catalog. Beyond the immediate exhibit - The system became a resource that visitors could return to after the initial interaction. 24

Design Implications Support observation, not only identification accuracy Allow partial and uncertain observations Support revisiting beyond the immediate exhibit experience Design for shared use in public spaces 25

Limitations & Future work • Limited generalizability: Single aquarium, one species, and a one-week deployment → Longer-term and multi-site studies with other species • Indirect measurement of engagement: Attention, learning, and internal engagement were not directly measured → Combine logs with questionnaires, interviews, and behavioral measures • User identification and self-selection: Browser-level IDs may not represent unique visitors, and participation was voluntary → Improve participant tracking and examine broader visitor populations 26

Thank you for listening. Paper Portfolio LinkedIn Contact : Yuki Nakagawa [email protected]