In this paper,we propose hierarchical attention dual network(DNet)for fine-grained image classification.The DNet can randomly select pairs of inputs from the dataset and compare the differences between them through hi...In this paper,we propose hierarchical attention dual network(DNet)for fine-grained image classification.The DNet can randomly select pairs of inputs from the dataset and compare the differences between them through hierarchical attention feature learning,which are used simultaneously to remove noise and retain salient features.In the loss function,it considers the losses of difference in paired images according to the intra-variance and inter-variance.In addition,we also collect the disaster scene dataset from remote sensing images and apply the proposed method to disaster scene classification,which contains complex scenes and multiple types of disasters.Compared to other methods,experimental results show that the DNet with hierarchical attention is robust to different datasets and performs better.展开更多
Two reagents including salicylhydroxamic acid(SHA) and tributyl phosphate(TBP) were tested as collectors either separately or together for electro-flotation of fine cassiterite(<10 μm).Subsequently,the flotation m...Two reagents including salicylhydroxamic acid(SHA) and tributyl phosphate(TBP) were tested as collectors either separately or together for electro-flotation of fine cassiterite(<10 μm).Subsequently,the flotation mechanism of the fine cassiterite was investigated by adsorbance determination,electrophoretic mobility measurements and Fourier transform infra-red(FT-IR) spectrum checking.Results of the flotation experiments show that with SHA as a collector,the collecting performance is remarkably impacted by the pulp pH value as the floatability of cassiterite varies sharply when the pH changes,and flotation with SHA gives distinct maximum at about pH 6.5.Additionally,the floatability of cassiterite is determined by using SHA and TBP as collectors.The range of pulp pH for good floatability is broadened in the presence of TBP as auxiliary collector,and the utilization of TBP improves the recovery of cassiterite modestly.Moreover,the optimum pH value for cassiterite flotation is associated with adsorbance.The results of FT-IR spectrum and the electrophoretic mobility measurements indicate that the adsorption interaction between the collectors and the cassiterite is dominantly a kind of chemical bonding in the form of one or two cycle chelate rings due to the coordination of carbonyl group,hydroxamate and P=O group to the metal tin atoms,where the oxygen atoms contained in carbonyl group,hydroxamate and P=O group of the polar groups have the stereo conditions to form five-membered rings.In addition,the adsorption interactions of SHA and TBP on the surfaces of cassiterite are also dominated by means of hydrogen bonds.展开更多
Deep learning has achieved excellent results in various tasks in the field of computer vision,especially in fine-grained visual categorization.It aims to distinguish the subordinate categories of the label-level categ...Deep learning has achieved excellent results in various tasks in the field of computer vision,especially in fine-grained visual categorization.It aims to distinguish the subordinate categories of the label-level categories.Due to high intra-class variances and high inter-class similarity,the fine-grained visual categorization is extremely challenging.This paper first briefly introduces and analyzes the related public datasets.After that,some of the latest methods are reviewed.Based on the feature types,the feature processing methods,and the overall structure used in the model,we divide them into three types of methods:methods based on general convolutional neural network(CNN)and strong supervision of parts,methods based on single feature processing,and meth-ods based on multiple feature processing.Most methods of the first type have a relatively simple structure,which is the result of the initial research.The methods of the other two types include models that have special structures and training processes,which are helpful to obtain discriminative features.We conduct a specific analysis on several methods with high accuracy on pub-lic datasets.In addition,we support that the focus of the future research is to solve the demand of existing methods for the large amount of the data and the computing power.In terms of tech-nology,the extraction of the subtle feature information with the burgeoning vision transformer(ViT)network is also an important research direction.展开更多
基金Supported by the National Natural Science Foundation of China(61601176)。
文摘In this paper,we propose hierarchical attention dual network(DNet)for fine-grained image classification.The DNet can randomly select pairs of inputs from the dataset and compare the differences between them through hierarchical attention feature learning,which are used simultaneously to remove noise and retain salient features.In the loss function,it considers the losses of difference in paired images according to the intra-variance and inter-variance.In addition,we also collect the disaster scene dataset from remote sensing images and apply the proposed method to disaster scene classification,which contains complex scenes and multiple types of disasters.Compared to other methods,experimental results show that the DNet with hierarchical attention is robust to different datasets and performs better.
基金Project(50774094) supported by the National Natural Science Foundation of ChinaProject(2010CB630905) supported by the National Basic Research Program of China
文摘Two reagents including salicylhydroxamic acid(SHA) and tributyl phosphate(TBP) were tested as collectors either separately or together for electro-flotation of fine cassiterite(<10 μm).Subsequently,the flotation mechanism of the fine cassiterite was investigated by adsorbance determination,electrophoretic mobility measurements and Fourier transform infra-red(FT-IR) spectrum checking.Results of the flotation experiments show that with SHA as a collector,the collecting performance is remarkably impacted by the pulp pH value as the floatability of cassiterite varies sharply when the pH changes,and flotation with SHA gives distinct maximum at about pH 6.5.Additionally,the floatability of cassiterite is determined by using SHA and TBP as collectors.The range of pulp pH for good floatability is broadened in the presence of TBP as auxiliary collector,and the utilization of TBP improves the recovery of cassiterite modestly.Moreover,the optimum pH value for cassiterite flotation is associated with adsorbance.The results of FT-IR spectrum and the electrophoretic mobility measurements indicate that the adsorption interaction between the collectors and the cassiterite is dominantly a kind of chemical bonding in the form of one or two cycle chelate rings due to the coordination of carbonyl group,hydroxamate and P=O group to the metal tin atoms,where the oxygen atoms contained in carbonyl group,hydroxamate and P=O group of the polar groups have the stereo conditions to form five-membered rings.In addition,the adsorption interactions of SHA and TBP on the surfaces of cassiterite are also dominated by means of hydrogen bonds.
基金supported by the National Natural Science Foundation of China(61571453,61806218).
文摘Deep learning has achieved excellent results in various tasks in the field of computer vision,especially in fine-grained visual categorization.It aims to distinguish the subordinate categories of the label-level categories.Due to high intra-class variances and high inter-class similarity,the fine-grained visual categorization is extremely challenging.This paper first briefly introduces and analyzes the related public datasets.After that,some of the latest methods are reviewed.Based on the feature types,the feature processing methods,and the overall structure used in the model,we divide them into three types of methods:methods based on general convolutional neural network(CNN)and strong supervision of parts,methods based on single feature processing,and meth-ods based on multiple feature processing.Most methods of the first type have a relatively simple structure,which is the result of the initial research.The methods of the other two types include models that have special structures and training processes,which are helpful to obtain discriminative features.We conduct a specific analysis on several methods with high accuracy on pub-lic datasets.In addition,we support that the focus of the future research is to solve the demand of existing methods for the large amount of the data and the computing power.In terms of tech-nology,the extraction of the subtle feature information with the burgeoning vision transformer(ViT)network is also an important research direction.
