Image Recognition Based on Mixed Attention Mechanism in Smart Home Appliances

Aiming at the problems of complexity of feature extraction and low recognition accuracy in the recognition process of oven food images, this paper proposes a method of oven food image recognition based on mixed attention mechanism. The model classifies and identifies the oven food samples by convolution neural network Xception, and the features can be extracted more quickly and effectively by the depth separable convolution in the network. However, there are some problems in the oven food image, such as small part of the food, complex background, background interference and so on, so the attention mechanism is added to the convolution neural network model. The attention mechanism gives different weights to different feature dimensions of the data, which realizes the enhancement of the food part in the oven food image and inhibits the irrelevant background. The experimental results show that the accuracy of this model for oven food recognition reaches 99 %, compared with the traditional convolution neural network has been greatly improved. The results confirm the effectiveness of the proposed identification method in this paper.