Spatial Perception

Spatial Perception helps assess the speed, stability, and accuracy of recognizing geometric shapes through four progressively more challenging stages: 1. Comparison of 2D shapes. 2. Comparison of 3D shapes. 3. Comparison of rotated 3D shapes. 4. Comparison of rotated 3D shapes with additional visual distractions. Tap “Same” when the shapes match or “Different” when they do not. Customise the test by choosing a visual texture (Chess, Grey Wood, or Color Abstraction), selecting 2, 3, or 4 stages, and setting 10-40 attempts per stage. You can explore 2D and 3D shapes, rotate 3D objects, and use Demo Test to preview all four stages. The Data section lets you review measurements, create personal reference models, and compare individual results with saved models, including the default Martial Arts model for highly qualified athletes. You can also save, import, replace, or delete data backups in JSON format. Scientific validation: The Spatial Perception assessment protocol and app demonstrated preliminary reliability and validity in a peer-reviewed study: Romanenko et al., 2025, Applied Sciences, 15(6), 3413. https://doi.org/10.3390/app15063413 The app includes an on-device Boosted Tree machine learning model (Core ML) that classifies four-stage test performance as High, Medium, or Low. It analyzes 12 stage-specific indicators: mean response time, response-time standard deviation, and error percentage for each stage. The model was trained on 278 combat sports athletes, validated on 60, and evaluated on a held-out test set of 59, achieving 91.5% testing accuracy. The app also calculates an integral SP Index based on overall response speed, response stability, and accuracy. A higher SP Index indicates faster, more stable, and more accurate performance. Results include the ML-predicted level and class probabilities, the index-based level, SP Index, an approximate reference percentile, and a score from 0 to 100. Spatial Perception can support regular performance monitoring by athletes, coaches, educators, and researchers. All calculations are performed locally on the device, and personal data are not transmitted to a server. Results should be interpreted in the context of age, preparedness, previous results, and testing conditions. The app does not predict competitive success and is not a medical or diagnostic device.