Artificial neural network adjustment method and device

Abstract

The invention provides an ANN (artificial neural network) adjustment method and device. The ANN includes a plurality of layers and is trained for classification reasoning. For example, the ANN may bea neural network trained for fine-grained image recognition. The method comprises the steps that a training data set is divided into a current category and other categories according to the currentlytargeted category, data features belonging to the current category are called as positive category features, and data features belonging to the other categories are called as negative category features; and the ANN is adjusted with a first loss function that specifies that the positive class feature is closer to a current class feature center of the current class than the negative class feature. According to the method, an effective loss function is designed from a loss function, the intra-class distance can be shortened, the inter-class distance can be increased, the classification effect canbe improved, and the prediction precision of the artificial neural network can be improved on the whole.

Description

technical field [0001] The invention relates to deep learning, in particular to an adjustment method and device for artificial neural networks. Background technique [0002] In recent years, artificial neural network (ANN) has made great progress in the fields of object detection and image classification. Among them, the fine-grained classification task is a challenging image recognition task to correctly identify objects from hundreds of subclasses of a base parent class. Since the difference between subclasses is usually small, and pose, rotation, etc. will cause large deviations within classes, it is difficult to improve the accuracy of fine-grained classification. In order to solve this problem, most of the current research relies on the localization of discriminative parts (such as bird's beaks, feet, etc.) Additional labor costs are added and the complexity of the system is increased. [0003] In view of this, there is still a need for an improved neural network tun...

AI Technical Summary

Problems solved by technology

Since the difference between subclasses is usually small, and pose, rotation, etc. will cause large deviations within the class, it is difficult to improve the accuracy of fine-grained classification

In order to solve this problem, most of the current research relies on the localization of discriminative parts (such as bird's beaks, feet, etc.) Additional labor costs and increased system complexity

Images