A spontaneous visual database for online learning We have uploaded some content of our dataset to the "Dataset" folder. What needs to be reminded is that due to privacy issues, we currently only provide extracted pixel values.
The folder contains 6 csv files, which save the pixel values of the facial images of the 6 academic emotions respectively. In each csv file, one row represents an image, the first column is the image name, and the following columns represent 48*48 pixel values. In emotion recognition tasks, features can be extracted based on pixel values, and pixel values can be directly used as the input of CNN.
In addition, we are currently processing other data in our database (including video clips and associated codes). After finishing, we will first release the recognition module data.