Multi-Input Convolutional Neural Network for Flower Grading

Figure 3

An architecture of 3-input CNN model for Gloxinia grading. This architecture has the best effect on testing set.