Figure 2:

Illustration of the architecture of the combined model for predicting pain progression. The proposed model consisted of two separate convolutional neural networks connected in a cascaded fashion to create a fully-automated pipeline. The combined model was created using YOLO and EfficientNet to extract DL information from baseline knee radiographs as a feature vector, which was further concatenated with the normalized demographic, clinical, and radiographic risk factor data vector. BN: batch normalization, Conv2D: 2D convolution, ReLU: rectified linear activation, 2D: two-dimensional.