Area of interest analysis

library(eyer)
obj <- eyer_data

Area allow us to define square areas and analyse the contents.

We can plot areas with the geom_area_boundaries function.

resolution <- obj$info$resolution
center <- AreaObject("center", c(resolution$width/2 - 200, resolution$width/2 + 200), 
                     c(resolution$height/2 - 200, resolution$height/2 + 200))
top_upper <- AreaObject("Top Upper", c(1500, 1920), c(750, 1080))
plot_gaze(obj, downsample = 50) + 
  geom_area_boundaries(center, alpha = 0.005, fill="blue") +
  geom_area_boundaries(top_upper, alpha = 0.005, fill="green") +
  geom_eyer_monitor(obj)

And then we can add information about which area is each recording is to the gaze and fixations

obj <- add_area_column(obj, list(center, top_upper))
str(obj$data$gaze)

## 'data.frame':    130000 obs. of  4 variables:
##  $ x   : num  1720 1720 1720 1721 1722 ...
##  $ y   : num  848 847 845 843 843 ...
##  $ time: int  0 1 2 3 4 5 6 7 8 9 ...
##  $ area: chr  "Top Upper" "Top Upper" "Top Upper" "Top Upper" ...

table(obj$data$gaze$area)

## 
##    center Top Upper 
##     33880     23861

Any recording not in given areas is assigned NA_character_. We can then select only those recordings in arease.

fixations_in_areas <- obj$data$fixations[!is.na(obj$data$fixations$area),]
ggplot(fixations_in_areas, aes(x,y, size=duration)) + geom_point() +
  geom_area_boundaries(center, alpha = 0.005, fill="blue") +
  geom_area_boundaries(top_upper, alpha = 0.005, fill="green") +
  geom_eyer_monitor(obj) + theme_minimal()

Analysis

df <- analyse_fixation_areas(obj, list(center, top_upper))
print(df)

##        area   n duration duration.na     ratio
## 1     total 378   114969           0 1.0000000
## 2 Top Upper  82    22110           0 0.1923127
## 3    center  99    30531           0 0.2655585

df <- analyse_gaze_areas(obj, list(center, top_upper))
print(df)

##        area      n     ratio
## 1     total 130000 1.0000000
## 2 Top Upper  23861 0.1835462
## 3    center  33880 0.2606154

Analysis

Contents