Fig. 4
Performance Metrics Across Subjects and Shapes (All Backgrounds). (a) F1-Score heatmap showing the harmonic mean of Precision and Recall for each subject and shape class, computed over trials including both white and black screen (dark) background conditions. (b) Confidence heatmap, displaying the average of Precision and Recall similarly across all background types. Values are derived from subject-wise classification data (Table 2), with color intensity indicating metric magnitude (scale: 0–0.8). Higher values reflect better model performance.
