Application of CNN deep learning in product design evaluation

Convolutional Neural Network (CNN) is an excellent deep learning algorithm. It can not only extract image features accurately, but also can reduce the complexity of the model. This paper combines the advanced technology of cognitive neurology, where we uses EEG equipment to read the real brain activity data when people make evaluation and at the same time we uses the eye-tracking equipment to collect the subject’s gaze point and gaze path, and then generates gaze hotspot map with gaze time. The CNN model is trained by the samples obtained by expert system scoring. Thanks to the advantages of CNN in image processing. AlexNet model with 8-layer network structure is used to extract the features of Brain Electrical Activity Mapping (BEAM) and gaze hot spot image, and then the Support Vector Machine (SVM) is used to classify and predict different degrees. Ultimately the product design features evaluation is achieved.