Searcharxiv⌕ Search

arXiv subjects

Bryceton Bible

Publications and source records attributed to Bryceton Bible.

3 recordsLinked to original sources

Small Data in Long-Term Environmental Vision: Annotation Noise, Gamification, and Label-Budget Scaling

Long-running environmental camera networks produce abundant images but few reliable annotations. Using FujiView, a network of webcams around Mount Fuji, we study three linked questions: whether gamification affects annotation participation and agreement, how annotators disagree on ordinal visibility labels, and how label budget affects downstream classification. Across seven nonrandomized sessions with 22 labelers, gamification showed a more promising pattern for sustained participation than for labeling speed or agreement. In the closest remote-session comparison, the session recorded as gamified had longer active labeling time, while throughput and reference agreement remained similar. Annotation disagreement was substantial at fine granularity: crowd votes matched the single-annotator reference on 49.1% of exact visibility classes, versus 90.2% after grouping them into Good/Poor. In five-replicate label-budget experiments, models using 25% of the available label pool achieved held-out exact accuracy only 2.8 percentage points below the full-pool model on average. Together, these findings highlight the importance of designing reliable annotation targets, evaluating engagement mechanisms without assuming causal benefits, and using empirical scaling curves to plan labeling effort.

cs.CV↗

Different Timescales, Different Signals: Multimodal Landmark-Visibility Prediction in a Long-Term Webcam Network

Scenic visibility is not a single forecasting problem: planning a trip months ahead, deciding whether to go outside today, and judging whether tomorrow will be clear rely on different information. Using a year-scale multimodal archive of Mount Fuji webcams and meteorological observations, we identify a clear hierarchy across timescales. Seasonal climatology provides useful long-range guidance, webcam imagery is the strongest same-day observation (AUC 0.95), and the numerical forecast valid on the target day carries most of the day-specific signal one to three days ahead. Revealing this hierarchy requires evaluation that respects the dependence among cameras observing the same regional weather state: permissive splits substantially overstate multi-day performance, while date-grouped and forward-in-time evaluation reduce the apparent contribution of the current scene and current weather. On 6,575 camera-days, the current visual summary provides no measurable one-day gain once the target-day forecast is available. The remaining opportunity lies at shorter horizons, where the present visual representation discards cloud geometry and motion. These results recast scenic visibility prediction as a multi-timescale sensing problem rather than a single multimodal forecasting task.

cs.CV↗

FujiView: Multimodal Late-Fusion for Predicting Scenic Visibility

Visibility of natural landmarks such as Mount Fuji is a defining factor in both tourism planning and visitor experience, yet it remains difficult to predict due to rapidly changing atmospheric conditions. We present FujiView, a multimodal learning framework and dataset for predicting scenic visibility by fusing webcam imagery with structured meteorological data. Our late-fusion approach combines image-derived class probabilities with numerical weather features to classify visibility into five categories. The dataset currently comprises over 100,000 webcam images paired with concurrent and forecasted weather conditions from more than 40 cameras around Mount Fuji, and continues to expand; it will be released to support further research in environmental forecasting. Experiments show that YOLO-based vision features dominate short-term horizons such as "nowcasting" and "samedaycasting", while weather-driven forecasts increasingly take over as the primary predictive signal beyond $+1$d. Late fusion consistently yields the highest overall accuracy, achieving ACC of approx 0.89 for same-day prediction and up to 84% for next-day forecasts. These results position Scenic Visibility Forecasting (SVF) as a new benchmark task for multimodal learning.

cs.CV↗