Search

The existing oil painting art style classification algorithms ignore the influence of the main area and the overall effect on the art style.Aiming at this problem,this paper proposes a new oil painting classification algorithm based on multi-feature fusion classifier(MFFC).Firstly,based on the common arrangement form of oil painting art elements,this paper designs the overlapping image block method.This method extracts spatial features of oil paintings to make up for the lack of composition style in existing algorithms.And it also can be used to distinguish the subject area from the background area.Secondly,the spatial features and the underlying features are combined in series to increase the location information of the elements in the picture.Finally,the spatial voting method is designed to give priority to the classification result of the main area as the output result of the algorithm.This is to highlight the role of oil painting subject area in the classification and realize the automatic classification of oil painting art style.Tested on the data set created by the FS-classifier model,its accuracy,precision,recall,F1-score and AUC reaches 96.92%,63.69%,98.75%,98.57% and 0.917,respectively.Compared with FS-classifier,the result increases by 6.72%,5.85%,9.05%,7.1% and 0.128,respectively.When tested on WIKIART and compared with other six algorithms,the accuracy improves by 13.27%,at least.The results show that the proposed algorithm can effectively improve the performance of spatial features for oil painting art style classification task,and has good practical value.

The technique of 3D deformation is one of the hot topics in the field of computer graphics.Current 3D deformation methods mainly learn the changes before and after deformation by aggregating localized adjacent voxel features,and fail to exploit the interrelationship between non-local voxel features,and the absence of contextual information prevents the model from capturing more discriminative features.To address the above problems,this paper designs a voxel deformation network based on environmental information mining,which can extract local and environmental information simultaneously,and extract environmental information from different spatial domains to improve the representation performance of the network,further modeling the relationship before and after the deformation of the object.Firstly,a novel self-attention mechanism is introduced.Specifically,the learning of the non-local dependence of different voxels is proposed to improve the ability of voxel discrimination.Then,a multi-scale analysis method is introduced to extract environmental information in different perceptual fields via multiple dilated convolution with different dilation rates,which provides more informative contextual features for the subsequent models.In addition,this paper analyzes the impact of feature fusion on the model and designs a method based on encoder-decoder feature fusion,which adaptively fuses the features extracted from the encoder and decoder to improve the nonlinear mapping capability of the model.Extensive experiments are conducted on our tooth dataset.The results show that the deformation prediction accuracy of the proposed method is improved compared to existing methods.

Different from artistic style transfer,the challenge of photorealistic style transfer is to maintain the authenticity of the output while transferring the color style of the style input.Now,most photorealistic style transfer methods perform pre-proces-sing or post-processing based on artistic style transfer methods,to maintain the authenticity of the output image.However,artistic style transfer methods usually cannot make full use of global color information to achieve a more coordinated overall impression,and pre-processing and post-processing operations are often tedious and time-consuming.To solve the above problems,this paper establishes a photorealistic style transfer network guided by global information,and proposes a color-partition-mean loss(Lcpm) to measure the similarity of the global color distribution between output and the style input.Adaptive instance normalization(AdaIN) is improved,and partition adaptive instance normalization(AdaIN-P) is proposed to better adapt to the color style transfer of real images.In addition,this paper also introduces a cross-channel partition attention module to make better use of global context information and improve the overall coordination of output images.Through the above methods,the decoder of network is guided to make full use of global information.Experimental results show that,compared with other state-of-the-art me-thods,the proposed model can achieve a better photorealistic style transfer effect while maintaining image details.

Person re-identification aims to retrieve specific pedestrian targets from the target database under the condition of cross camera.It has important application value in the field of video surveillance.At present,the difficulty of the research is that the sample images have large intra class differences and small inter class differences.Therefore,how to design and train the deep neural network to extract a more discriminative feature from pedestrian images is the key.In this paper,we propose a network structure combining global features and local features learning,which can extract global features and local features simultaneously.In view of the different importance of each part of the local features to the description of pedestrian features,this paper proposes a fusion method of local features,which can adaptively generate the weight of each local feature.Finally,the local features and glo-bal features are combined to make the pedestrian features get more comprehensive representation.In addition,in view of the fuzzy optimization objective of the previous triple loss based on hard sample mining,this paper proposes an improved triple loss function based on hard sample mining.The effectiveness of the proposed method is verified on the mainstream person re-identification data sets Market-1501 and DukeMTMC-reID,respectively,and the mAP values are 82.16% and 74.02%,and the Rank-1 values are 92.75% and 86.8%,respectively.

Few-shot object detection aims to train target detection model through a small amount of sample learning.At present,most of the existing few-shot object detection methods are based on classical target detection algorithms.In the two-stage detection method,due to the small number of new class samples,many irrelevant border boxes are generated,resulting in low accuracy of candidate regions.To solve this problem,this paper proposes a few-shot object detection algorithm FF-FSOD based on feature fusion.It uses the feature fusion method to enhance the data,supplements the new category samples,increases the coverage range of the sample,and introduces the FPN network to extract multi-scale feature.Then,the RPN network is improved,and the support set image branch is introduced.The depth correlation between the support set image feature and the query set image feature is calculated,and the attention feature map is obtained,and the more accurate candidate box is obtained.The effectiveness of the proposed model is verified on MS COCO and FSOD datasets.Experimental results show that the proposed method obtains more accurate candidate boxes and improves the detection accuracy.

The online judge (OJ) is a widely used system for programming education,learning and contests.Users often get lost in searching for tasks of interest in the massive database.How to recommend suitable programming tasks to the users and plan the learning path is a significant research topicin the development of online programming evaluation system.Existing traditional recommendation methods have the limitation of making a trade-off between interpretability and effectiveness.This paper proposes a task-recommending model for the OJ platform -hybrid programming task recommendation model based on knowledge graph and collaborative filtering for online judge (HKGCF).The HKGCF model can help users improve their learning effect by recommending questions that match their current knowledge levels and skills.The model is designed based on a hybrid strategy that integrates the knowledge graph representation learning with an improved collaborative filtering algorithm.The model is implemented and integrated into the OJ platform of Beihang University,and meet the specific interaction formats of the OJ platform.We conducted two experiments,an online and an offline test,to validate the proposed model and its implementations.The results show that the proposed model outperforms the representative conventional recommendation algorithm interms of interpretability and accuracy

The imbalanced problem is common in the real world,and the highly-skewed distribution of imbalanced data seriously affects the performance of the model.In general,the imbalanced data affects the model performance from two aspects.On the one hand,the imbalance in sample size leads to more updates of parameters in majority classes,which leads to the model biased to majority classes.On the other hand,the sample size of minority classes is too small,and the diversity is insufficient,which leads to the insufficient representation ability of the model.To solve these problems,this paper proposes a feature fusion framework combining attention mechanism and geometric information.Specifically,in the first stage,the model learns the semantic information and discriminative information of the data through pre-training,and combines the attention mechanism to discover where the mo-del pays more attention.In the second stage,the model uses geometric information to mine boundary features,and combines the attention weight obtained in the first stage to fuse the boundary features,so as to supplement minority classes.Experimental results on long tail CIFAR10,CIFAR100 and KDD Cup99 datasets show that the proposed feature fusion framework combining attention mechanism and geometric information can effectively improve the classification performance of imbalanced data,and can effectively improve the classification performance for different types of data,including image data and structured data.

This paper proposes a reflective vest detection algorithm for apron staff based on prior knowledge and improved YOLOv3 algorithm.Aiming at the problem of the existing target detection method with low speed, the reflective vest detection candidate region is generated based on prior knowledge to replace the initial candidate region, so as to reduce the detection area.Darknet-37 is used to replace Darknet-53 as the backbone network for feature extraction, which improves the detection speed of the algorithm.Aiming at the problem that the reflective vest occupies a small area in the picture and is difficult to identify, a spatial pyramid pooling structure (SPP) is added into the detection model to realize feature enhancement, and the detection scale is increased to four for multi-scale feature fusion.The K-means++algorithm is used to perform cluster analysis again on the size of labeled bounding box, and the clustering result is used to replace the initial Anchor value of Yolov3.GIoU is selected as the loss function to improve the positioning accuracy.Experimental results show that the proposed new target detection algorithm in the self-built reflective vest data set is better than YOLOv3 test results, the precision rate and recall rate reach 97.6% and 96.1%, detection rate reach 28.4 frames/s, which effectively solves the problems such as inaccurate positioning, missed detection and low detection speed existing in the original model, and meets the real-time requirements in the practical application of apron target detection while ensuring a high detection accuracy.

Pedestrian appearance attributes are important semantic information distinguishing pedestrian differences.Pedestrian attribute recognition plays a vital role in intelligent video surveillance,which can help us quickly screen and retrieve target pedestrians.In the task of person re-identification,we can use attribute information to obtain fine feature expressions,thereby improving the effect of pedestrian re-identification.This paper attempts to combine pedestrian attribute recognition with person re-identification,looking for a way to improve the performance of person re-identification,and proposes a person re-identification framework based on feature positioning and fusion.Firstly,we use the method of multi-task learning to combine person re-identification with attribute recognition,and improve the performance of the network model by modifying the convolution step size and using double pooling.Secondly,to improve the expression ability of attribute features,a parallel spatial channel attention module based on the attention mechanism is designed.It can not only locate the spatial position of the attribute on the feature map,but also can effectively mine the channel with higher correlation with the attribute features,and uses multiple groups of parallel branch structure to reduce errors and further improve the performance of the network model.Finally,we use the convolutional neural network to design the feature fusion module to effectively integrate the attribute features and pedestrian identity features to obtain more robust and expressive pedestrian features.The experiment is conducted on two commonly used person re-identification datasets DukeMTMC-reID and Market-1501.The results show that this method is at the leading level among the existing person re-identification methods.

With the rapid development of mobile internet,social network platform is full of a large number of text data with emotional color.Mining the emotion in such text not only helps to understand the attitude and emotion of internet users,but also plays an important role in scientific research institutions and the government to grasp the emotional changes and trends of society.Traditional sentiment analysis mainly focuses on the analysis of sentiment tendency,which can not accurately and multi-dimensionally describe the emotion of the text.In order to solve this problem,this paper studies the emotion analysis of the text.Firstly,aiming at the lack of fine-grained sentiment tags in text data sets of different fields,a deep learning based emotion analysis model,FMRo-BLA,is proposed.The model pre-trains the general domain text,and then applies the pre-trained model to the downstream situation of specific domain through parameter based migration learning,feature fusion and FGM Adversarial training.Compared with the best performance of RoBERTa pre-trained language model,the F1 value of the proposed method is improved by 5.93% on the target domain dataset,and it achieves 12.38% by further adding transfer learning.

Micro-expression refers to an uncontrollable muscle movement on the face when people try to hide or suppress their true emotions.Due to the short duration,small motion range,and difficulty in concealing and restraining,the recognition accuracy of such emotional facial expressions is restricted.In order to cope with these challenges,this paper proposes a novel micro-expression recognition method combining feature fusion and attention mechanism,considering optical flow features and face features,and further adding attention mechanism to improve the recognition performance.The processing steps of this method are as follows:1)Extract the optical flow and optical strain from Onset to Apex in each micro-expression segment,input the vertical optical flow,horizontal optical flow and optical strain into a shallow 3DCNN,and extract the optical flow features.2)Taking the deep convolution neural network ResNet-10 as the backbone network,the convolution attention module is added to extract face features.3)Combine the two feature vectors for classification.The experimental results reveal that the proposed method is superior to the traditional methods and existing deep learning methods in micro-expression recognition.

Due to the high spatial resolution of high-resolution remote sensing images,rich ground objects information,high complexity,uneven distribution of target categories and different sizes of various ground objects,it is difficult to improve the segmentation accuracy.In order to improve the semantic segmentation accuracy of remote sensing images and solve the problem that U-Net model is limited when combining deep semantic information and shallow position information,a semantic segmentation me-thod of remote sensing images based on U-Net feature fusion optimization strategy is proposed.This method adopts the encoder-decoder structure based on U-Net network.In the feature extraction part of the network,the encoder structure of U-Net model is used to extract the feature information of multiple layers.In the feature fusion part,the jump connection structure of U-Net is retained,and at the same time,the feature fusion optimization strategy proposed in this paper is used to realize the fusion-optimization-refusion of high-level semantic features and low-level location features.In addition,the feature fusion optimization strategy uses dilated convolution to get more global features,and uses Sub-Pixel convolutional layer instead of traditional transposed convolution to achieve adaptive upsampling.This method is validated on the Potsdam dataset and Vaihingen dataset of ISPRS.The three evaluation indexes,overall classification accuracy,Kappa coefficient and mIoU in the verification are 86.2%,0.82,0.77 on Potsdam dataset,and 84.5%,0.79,0.69 on Vaihingen dataset.Compared with the traditional U-Net model,the three evaluation indicators are increased by 5.8%,8%,8% on Potsdam dataset,and 3.5%,4%,11% on Vaihingen dataset.Experimental results show that the remote sensing image semantic segmentation method based on the U-Net feature fusion optimization strategy has achieved good semantic segmentation effects on both the Potsdam dataset and the Vaihingen dataset,which can improve the accuracy of semantic segmentation of remote sensing images.

Choroidal neovascularization (CNV) generally occurs at the late stage of senile macular degeneration (AMD),and accurate segmentation of CNV in optical coherence tomography (SD-OCT) is of great significance for the diagnosis and treatment of AMD.This paper proposes a CNV multi-task segmentation network that combines time series model and attention mechanism.The continuous SD-OCT image is input into the segmentation network,and the multi-scale information of the picture is extracted in the encoder part.In order to better extract the local features of the picture,the attention gate is added in the skip connection part.In order to solve the problem of discontinuous scanning segmentation,after the segmentation network is pooled,the timing constraint network is passed to generate the continuity constraint of adjacent frames and gradient constraints are added to the loss function to better preserve the lesion boundary.The spatial pyramid is used to fuse the two parts of the network feature map to produce segmentation loss,which improves the final segmentation accuracy.Based on patient independence,effective cross-validation is performed on 200 eyes of 12 patients.The Dice coefficient reaches 76.3% and the overlap reaches 60.7%.CNV can be reliably segmented in SD-OCT images.