Skip to main content
. 2023 Sep 18;44(5):967–980. doi: 10.24272/j.issn.2095-8137.2022.449

Figure 1.

Figure 1

Examples of MiL dataset, with videos corresponding to actions and postures (labels)

A–J: Ten action categories. A: Climb; B: Hang; C: Turn; D: Walk; E: Shake; F: Jump; G: Move down; H: Lie down; I: Stand up; J: Sit down. K–M: Three posture categories. K: Stand; L: Sit; M: Huddle. Each row represents non-contiguous frames randomly sampled in the corresponding video. Video lengths range from 20 to 110 frames.