ABNORMAL OBJECT RECOGNITION METHOD AND APPARATUS, MEDIUM, AND ELECTRONIC DEVICE

摘要

The present application relates to the field of neural networks, and provides an abnormal object recognition method and apparatus, a medium, and an electronic device. The method comprises: obtaining object data and labels which correspond to the object data and represent whether objects are abnormal; dividing the object data into a training set and a test set; inputting the object data in the training set and corresponding labels to multiple deep neural network models to be trained for training to obtain multiple models; inputting the object data in the test set into the deep neural network models to obtain abnormal probabilities output by the models; determining a target deep neural network model according to the abnormal probabilities output by the models; cascading the target deep neural network model with an extreme gradient boosting model to obtain a cascade model, and training the cascade model using the training set to obtain a trained cascade model; and inputting object data to be recognized into the trained cascade model for prediction. The method improves the abnormal object recognition accuracy and reduces the missing recognition rate of abnormal objects.