Fine-Grained Image Retrieval via Piecewise Cross Entropy loss

Zeng, Xianxian; Zhang, Yun; Wang, Xiao Dong; Chen, Kairui; Liu, Dong; Yang, Weijun

doi:10.1016/j.imavis.2019.10.006

Cited by 34 publications

(29 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…e centralized ranking loss [2] proposed by Zheng et al in 2018 and the decor related centralized loss [33] proposed in 2019 combine the loss with weakly supervised feature aggregation to achieve better retrieval accuracy. A large number of subsequent finegrained image retrieval research mainly focused on the method of deep metric learning [35,36,40]. is method trains deep convolutional neural networks by setting a specific loss function to make a more ideal embedding space.…”

Section: Fine-grained Image Retrieval (Fgir)mentioning

confidence: 99%

“…Proxy-based losses include proxy nca loss [45], center loss [23], decorrelated centralized loss [33], and piecewise cross-entropy loss [36]. Proxy-based loss introduces the concept of the global proxy so that all classes of the training set can be considered while reducing the amount of calculation instead of being limited to samples in a minibatch.…”

Section: Scientific Programmingmentioning

confidence: 99%

“…(b) e previous loss usually only considers the similarity between each proxy and each sample, or the similarity between each image sample obtained after convolution, but it does not consider the similarity between each proxy and each proxy [31][32][33]. And distributing a single proxy to each class cannot sufficiently characterize the intraclass changes of finegrained images [34][35][36]. (c) For fine-grained retrieval problems, deep convolution features are better than those of fully connected layers [3,25,[37][38][39].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multiproxies Adaptive Distribution Loss with Weakly Supervised Feature Aggregation for Fine-Grained Retrieval

Zhao

Zhang

2021

Scientific Programming

View full text Add to dashboard Cite

Fine-grained retrieval is one of the complex problems in computer vision. Compared with general content-based image retrieval, fine-grained image retrieval faces more difficult challenges. In fine-grained image retrieval tasks, all classes belong to a subclass of a meta-class, so there will be small interclass variance and large intraclass variance. In order to solve this problem, in this paper, we propose a fine-grained retrieval method to improve loss and feature aggregation, which can achieve better retrieval results under a unified framework. Firstly, we propose a novel multiproxies adaptive distribution loss which can better characterize the intraclass variations and the degree of dispersion of each cluster center. Secondly, we propose a weakly supervised feature aggregation method based on channel weighting, which distinguishes the importance of different feature channels to obtain more representative image feature descriptors. We verify the performance of our proposed method on the universal benchmark datasets such as CUB200-2011 and Stanford Dog. Higher Recall@K demonstrates the advantage of our proposed method over the state of the art.

show abstract

Section: Fine-grained Image Retrieval (Fgir)mentioning

confidence: 99%

Section: Scientific Programmingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multiproxies Adaptive Distribution Loss with Weakly Supervised Feature Aggregation for Fine-Grained Retrieval

Zhao

Zhang

2021

Scientific Programming

View full text Add to dashboard Cite

show abstract

“…e cross-entropy loss function is often applied for classification problems, especially for the classification problem in neural networks [30], and the cross entropy is used as the loss function frequently. In addition, since cross-entropy involves calculating the probability of each category, cross entropy appears with the Sigmoid (or softmax) function [31] almost every time. e expression of the Sigmoid function is as follows:…”

Section: Input Layer Convolutional Layermentioning

confidence: 99%

Chinese Microblog Sentiment Detection Based on CNN-BiGRU and Multihead Attention Mechanism

Qiuhong¹,

FanChongdi

YaoJie³

et al. 2020

Scientific Programming

View full text Add to dashboard Cite

With the rapid development of the Internet, Weibo has gradually become one of the commonly used social tools in society at present. We can express our opinions on Weibo anytime and anywhere. Weibo is widely used and people can express themselves freely on it; thus, the amount of comments on Weibo has become extremely large. In order to count up the attitudes of users towards a certain event, Weibo managers often need to evaluate the position of a certain microblog in an appropriate way. In traditional position detection tasks, researchers mainly mine text semantic features through constructing feature engineering and sentiment dictionary, but it takes a large amount of manpower in feature selection and design. However, it is an effective method to analyze the sentiment state of microblog comments. Deep learning is developing in an increasingly mature direction, and the utilization of deep learning methods for sentiment detection has become increasingly popular. The application of convolutional neural networks (CNN), bidirectional GRU (BiGRU), and multihead attention mechanism- (multihead attention-) combined method CNN-BiGRU-MAttention (CBMA) to conduct Chinese microblog sentiment detection was proposed in this paper. Firstly, CNN were applied to extract local features of text vectors. Afterward, BiGRU networks were applied to extract the global features of the text to solve the problem that the single CNN cannot obtain global semantic information and the disappearance of the traditional recurrent neural network (RNN) gradient. At last, it was concluded that the CBMA algorithm is more accurate for Chinese microblog sentiment detection through a variety of algorithm experiments.

show abstract

“…Deep neural networks (DNNs) based on the fully convolutional neural network have showed great improvements over systems relying on hand-crafted features [1][2][3] on benchmark tasks. With the rapid progress in DNNs research in recent years, it has dramatically facilitated the development of computer vision, such as object detection [4][5][6], image retrieval [7][8][9], scene recognition [10,11], semantic segmentation [12][13][14], image classification and inpainting [15,16], and so on. In particular, the state-of-theart works in object detection continues to grow, including face recognition [17][18][19], pedestrian detection [20][21][22], vehicle detection [23,24], etc.…”

Section: Introductionmentioning

confidence: 99%

Pedestrian as Points: An Improved Anchor-Free Method for Center-Based Pedestrian Detection

et al. 2020

View full text Add to dashboard Cite

Fine-Grained Image Retrieval via Piecewise Cross Entropy loss

Cited by 34 publications

References 5 publications

Multiproxies Adaptive Distribution Loss with Weakly Supervised Feature Aggregation for Fine-Grained Retrieval

Multiproxies Adaptive Distribution Loss with Weakly Supervised Feature Aggregation for Fine-Grained Retrieval

Chinese Microblog Sentiment Detection Based on CNN-BiGRU and Multihead Attention Mechanism

Pedestrian as Points: An Improved Anchor-Free Method for Center-Based Pedestrian Detection

Contact Info

Product

Resources

About