Gap Loss for Production Status Detection

Abstract

In order to classify the monitoring model of the production state system more accurately and effectively, and accelerate the convergence speed of the model, a new loss function Gap Loss function is proposed. The image data and sensor data collected by the system in real time are used as feature data, the five self-defined states are the learning targets, and Gap Loss is used instead of the cross-entropy loss function for model training. Experimental verification shows that this method can more effectively classify production states compared with the cross entropy Loss function.