Skip to main content
. 2022 Aug 23;11:e74314. doi: 10.7554/eLife.74314

Figure 10. Comparisons with JAABA.

Figure 10.

(A) MoTr tracking in a standard fear conditioning chamber (left) and an open field (right). (B) Ctrax tracking in a standard fear conditioning chamber (left) and an open field (right). (C) DLC tracking in a standard fear conditioning chamber (left) and estimated ellipse based on keypoint tracking (right). (D) Quantification of freezing detection errors for BehaviorDEPOT and JAABA (Precision: FClassifier(1,12)=0.58, P=0.46; Recall: FClassifier(1,11)=51.27, P<0.001; F1: FClassifier(1,11)=51.27, P<0.001, N=4–6 videos, 2-way ANOVA with Sidak’s multiple comparison test). (E) Comparison of VTE detection by human, BehaviorDEPOT, and JAABA (FClassifier(1.072,2.143)=340.2, P=0.0021, N=3 videos, Repeated measures one-way ANOVA with Tukey multiple comparison’s test). Scale bars, 5 cm.