Adult jerboas spend the majority of their time with food-directed behaviors.

相似書籍

• Scoring and quantification of jerboa alopecia.
  由: Matthew D. Boulanger (22600966)
  出版: (2025)
• The majority of jerboa behaviors are rapid, and the duration of behaviors significantly influences observer and classifier annotation accuracy.
  由: Matthew D. Boulanger (22600966)
  出版: (2025)
• Machine classifiers successfully met human-grade performance standards, including both accuracy and precision. a) After orientation, human observers (n = 11, orange) annotated jerboa behavior in the comparison subset. Adequately trained independent machine classifiers (n = 11, blue) also annotated the same comparison subset (AM = Ambulating, WI = Wall Interaction, RM = Rapid, Upright Movements, FB = Food Bowl Interactions, JU = Jumping, DB = Dust Bath Interactions, GR = Grooming, RO = Rolling.) b) This resulted in human observers and machine classifiers achieving mean accuracies of 0.71 ± 0.11 (orange line) and 0.76 ± 0.069 (blue line), respectively. Statistically significant differences between these means were not noted (p > 0.05). c) Similarly, human observers and machine classifiers achieved mean ICC values of 0
  由: Matthew D. Boulanger (22600966)
  出版: (2025)
• The fully trained computer vision classifier was 0.78 accurate overall, according to internal metrics [17], but accuracy varied among behaviors. Notably, machine classification of jumping and grooming was more accurate than human observers (Fig 5).
  由: Matthew D. Boulanger (22600966)
  出版: (2025)
• The incidence of grooming is most influenced by the addition of an opaque divider, as compared to manipulanda.
  由: Matthew D. Boulanger (22600966)
  出版: (2025)