Figure 2 | Scientific Reports

Figure 2

From: Detecting medial patellar luxation with ensemble deep convolutional neural network based on a single rear view image of the hindlimb

Figure 2

Architecture of the proposed Ensemble model based on ResNet18. Our proposed model consists of (i) a dog-specific discriminator, (ii) a dog-specific encoder (iii) an ensemble classifier. Specifically, a ResNet18-based dog-specific encoder consisted of an input layer, and 18 convolutional neural network layers with ReLU activation functions in five residual blocks with adaptive-pooling layers, an FC layer, and sigmoid. The final sigmoid activation function has been used to predict normal or Medial Patellar Luxation (MPL). (2-Column).

Back to article page