Annotated Facial Landmarks for attention and inattention

This data consists of 34 facial landmark features of 18 children with autism spectrum disorder and 26 typically developing children between ages 7 and 11 years old during attention tasks experiment,  The attention task consists of 4 levels where random letters are displayed on the board presented in a virtual classroom: level 1(none: without distractions). Other observation samples were generated from level 2-4 (easy, medium, and hard: with distractions).

iMotions software detects to generate the facial landmarks from five regions of the face: eyes, eyebrows, nose, lips, and jaw. Two class labels— Attention and Inattention were generated based on task performance and some attention behavioral rules, such as looking at the target letters.

The video stream data generated during the attention task is divided into separate time slots for each random letter displayed. Each letter appears for the duration of 1400ms. Then each slot was labeled as attention and inattention for correct (X-click) and incorrect click (Missed X), respectively. The total observation samples were extracted from 95 videos of all the participants. Each length of the video was 300 seconds long and iMotions software reads the video stream at 16 frame rate per second. The frames were annotated for attention and inattention when the participants responded to the target letter and iMotions generated 16 data samples in1 second.