Combining biological motion perception with optic flow analysis for self-motion in crowds

Abstract

Heading estimation from optic flow relies on the assumption that the visual world is rigid. This assumption is violated when one moves through a crowd of people, a common and socially important situation. The motion of people in the crowd contains cues to their translation in the form of the articulation of their limbs, known as biological motion. We investigated how translation and articulation of biological motion influence heading estimation from optic flow for self-motion in a crowd. Participants had to estimate their heading during simulated self-motion toward a group of walkers who collectively walked in a single direction. We found that the natural combination of translation and articulation produces surprisingly small heading errors. In contrast, experimental conditions that either present only translation or only articulation produced strong idiosyncratic biases. The individual biases explained well the variance in the natural combination. A second experiment showed that the benefit of articulation and the bias produced by articulation were specific to biological motion. An analysis of the differences in biases between conditions and participants showed that different perceptual mechanisms contribute to heading perception in crowds. We suggest that coherent group motion affects the reference frame of heading perception from optic flow.

Figure 1.

Single frame of the stimulus with a crowd of forwards moving point-light walkers. The stimulus consisted of a group of point-light walkers walking coherently into a common direction simultaneously with simulated forward movement of the observer. The white arrow (left) indicates the direction of movement of the point-light walkers in this example. The red arrow (right) indicates the simulated self-motion of the observer.

Figure 2.

Median unsigned heading errors across all subjects and facing directions for each experimental condition. Error bars give the upper and lower interquartile ranges.

Figure 3.

Median signed heading errors in the natural articulation-plus-translation condition for facing directions near the straight ahead (0°). Positive values indicate a heading bias to the right and negative values indicate a heading bias to the left. Error bars give the upper and lower interquartile ranges.

Figure 4.

Median signed heading errors in the natural articulation-plus-translation condition for all facing directions. Error bars give the upper and lower interquartile ranges.

Figure 5.

Median unsigned heading errors in the natural articulation-plus-translation condition for all facing directions. Error bars give the upper and lower interquartile ranges.

Figure 6.

Signed heading errors of all individual participants in the natural articulation-plus-translation condition. Points represent individual means over all trials for a particular facing. Curves present local regressions through individual the data. Participants are color coded.

Figure 7.

Signed heading errors of individual participants in the only-translation condition. According to the direction of bias in this condition participants were separated into two groups. (a) Participants (n = 17) of group A show a heading bias against the direction of group motion, that is, positive and to the right for leftward (negative) facing directions and negative and to the left for rightward (positive) facing directions. The thick black line shows a prediction if participants simply responded as if all motion was due to self-motion, and hence, reported the vector average of the true heading and the inverse of the translation of the group. (b) Participants (n = three) of group B instead displayed an opposite behavior, that is, a bias in the direction of crowd motion.

Figure 8.

Signed heading errors of individual participants from the two groups of Figure 7 in the only-articulation condition. (a) Those individuals who showed a bias against the facing direction in the only-translation condition (group A). (b) Those individuals who showed a bias in the facing direction in the only-translation condition (group B).

Figure 9.

Relation of signed heading errors between the only-translation and only-articulation conditions in individual data. Each color represents an individual participant. (a) Scatterplot of individual data from group A (n = 17) with corresponding linear regressions. Thirteen of 17 participants displayed a negative correlation (blue/green colors) and four participants displayed a weak to moderate positive correlation (yellow colors). (b) Scatterplot of individual data from group B (n = three) with corresponding linear regressions. All three participants had a strong positive correlation of heading error between conditions.

Figure 10.

Results of Experiment 2. (a, b) Mean errors in facing direction for the normal and nonbiological walker types in the natural articulation-plus-translation condition (a) and the only-articulation condition (b). Error bars give the standard deviations of the mean. (c) Errors in facing direction in the only-articulation condition from all individual participants.