“…Image and video classifications remain the focus of cutting-edge research where deep learning convolutional neural networks can accomplish tasks in hours more accurately than would take humans days and weeks (He, Zhang, Ren, & Sun, 2016;Kwak & An, 2016;Sermanet et al, 2013). As researchers have increasing access to off-the-shelf or open-source libraries such as ImageNet, PASCAL VOC, and TensorFlow to facilitate image analysis in quick timeframes for additional analysis (Abadi, Isard, & Murray, 2017;Klostermann et al, 2018;Krizhevsky, Sutskever, & Hinton, 2012), these methods can be applied more widely by more substantively focused researchers. While current work has become proficient at object recognition in visual data (You, Luo, Jin, & Yang, 2015), many conventional approaches still have difficulties recognizing and extracting more abstract information such as emotions and sentiments, which is admittedly tricky even for humans coders due to subjectivity (Wang & Li, 2015).…”