Visual attentional-driven deep learning method for flower recognition

Abstract

<abstract> <p>As a typical fine-grained image recognition task, flower category recognition is one of the most popular research topics in the field of computer vision and forestry informatization. Although the image recognition method based on Deep Convolutional Neural Network (DCNNs) has achieved acceptable performance on natural scene image, there are still shortcomings such as lack of training samples, intra-class similarity and low accuracy in flowers category recognition. In this paper, we study deep learning-based flowers' category recognition problem, and propose a novel attention-driven deep learning model to solve it. Specifically, since training the deep learning model usually requires massive training samples, we perform image augmentation for the training sample by using image rotation and cropping. The augmented images and the original image are merged as a training set. Then, inspired by the mechanism of human visual attention, we propose a visual attention-driven deep residual neural network, which is composed of multiple weighted visual attention learning blocks. Each visual attention learning block is composed by a residual connection and an attention connection to enhance the learning ability and discriminating ability of the whole network. Finally, the model is training in the fusion training set and recognize flowers in the testing set. We verify the performance of our new method on public Flowers 17 dataset and it achieves the recognition accuracy of 85.7%.</p> </abstract>

Using Channel and Network Layer Pruning Based on Deep Learning for Real-Time Detection of Ginger Images

“…This study constructs a ginger recognition network based on YOLO v3 and pre-trains the Darknet-53 model using the ImageNet [34] dataset. Then, its first 74 layers weights are used as the pre-training model to achieve transfer learning [35][36][37]. (3) Ginger recognition model optimization and pruning.…”

Swin Routiformer: Moss Classification Algorithm Based on Swin Transformer With Bi-Level Routing Attention

“…Xu Jinghui et al [11] employed transfer learning to train the VGG16 model, achieving recognition of diseases in corn leaves, such as maize leaf spot and rust, with an accuracy as high as 95.33% through rotations, flips, and other data augmentation techniques. Cao Shuai et al [12] proposed a novel visual attentiondriven deep residual network that significantly improved the accuracy of flower recognition tasks across different channel features and spatial dimensions without introducing additional training parameters. In the moss classification task, Takeshi et al [13] introduced a method called "chunking" to process moss image data and used LeNet to classify three types of mosses with an accuracy exceeding 90%.…”