Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware reso...Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.展开更多
Watershed segmentation is sensitive to noises and irregular details within the image,which frequently leads to a serious over-segmentation Linear filtering before watershed segmentation can reduce over-segmentation to...Watershed segmentation is sensitive to noises and irregular details within the image,which frequently leads to a serious over-segmentation Linear filtering before watershed segmentation can reduce over-segmentation to some extent,however,it often causes the position offset of object contours.For the purpose of reducing over-segmentation to preserve the location of object contours,the watershed segmentation based on the hierarchical multi-scale modification of morphological gradient is proposed.Firstly,multi-scale morphological filtering was employed to smooth the original image.Then,the gradient image was divided into multi-levels by the volume of three-dimension topographic relief,where the lower gradient layers were further modifiedby morphological closing with larger-sized structuring-elements,and the higher layers with the smaller one.In this way,most local minimums caused by irregular details and noises can be removed,while region contour positions corresponding to the target area were largely preserved.Finally,morphological watershed algorithm was employed to implement segmentation on the modified gradient image.The experimental results show that the proposed method can greatly reduce the over-segmentation of the watershed and avoid the position offset of the object contours.展开更多
Grasp detection plays a critical role for robot manipulation.Mainstream pixel-wise grasp detection networks with encoder-decoder structure receive much attention due to good accuracy and efficiency.However,they usuall...Grasp detection plays a critical role for robot manipulation.Mainstream pixel-wise grasp detection networks with encoder-decoder structure receive much attention due to good accuracy and efficiency.However,they usually transmit the high-level feature in the encoder to the decoder,and low-level features are neglected.It is noted that low-level features contain abundant detail information,and how to fully exploit low-level features remains unsolved.Meanwhile,the channel information in high-level feature is also not well mined.Inevitably,the performance of grasp detection is degraded.To solve these problems,we propose a grasp detection network with hierarchical multi-scale feature fusion and inverted shuffle residual.Both low-level and high-level features in the encoder are firstly fused by the designed skip connections with attention module,and the fused information is then propagated to corresponding layers of the decoder for in-depth feature fusion.Such a hierarchical fusion guarantees the quality of grasp prediction.Furthermore,an inverted shuffle residual module is created,where the high-level feature from encoder is split in channel and the resultant split features are processed in their respective branches.By such differentiation processing,more high-dimensional channel information is kept,which enhances the representation ability of the network.Besides,an information enhancement module is added before the encoder to reinforce input information.The proposed method attains 98.9%and 97.8%in image-wise and object-wise accuracy on the Cornell grasping dataset,respectively,and the experimental results verify the effectiveness of the method.展开更多
In order to improve the detection accuracy of small objects,a neighborhood fusion-based hierarchical parallel feature pyramid network(NFPN)is proposed.Unlike the layer-by-layer structure adopted in the feature pyramid...In order to improve the detection accuracy of small objects,a neighborhood fusion-based hierarchical parallel feature pyramid network(NFPN)is proposed.Unlike the layer-by-layer structure adopted in the feature pyramid network(FPN)and deconvolutional single shot detector(DSSD),where the bottom layer of the feature pyramid network relies on the top layer,NFPN builds the feature pyramid network with no connections between the upper and lower layers.That is,it only fuses shallow features on similar scales.NFPN is highly portable and can be embedded in many models to further boost performance.Extensive experiments on PASCAL VOC 2007,2012,and COCO datasets demonstrate that the NFPN-based SSD without intricate tricks can exceed the DSSD model in terms of detection accuracy and inference speed,especially for small objects,e.g.,4%to 5%higher mAP(mean average precision)than SSD,and 2%to 3%higher mAP than DSSD.On VOC 2007 test set,the NFPN-based SSD with 300×300 input reaches 79.4%mAP at 34.6 frame/s,and the mAP can raise to 82.9%after using the multi-scale testing strategy.展开更多
A concept map is a diagram depicting relationships among concepts which is used as a knowledge representation tool in many knowledge domains. In this paper, we build on the modeling framework of Hui et al. (2008) in o...A concept map is a diagram depicting relationships among concepts which is used as a knowledge representation tool in many knowledge domains. In this paper, we build on the modeling framework of Hui et al. (2008) in order to develop a concept map suitable for testing the empirical evidence of theories. We identify a theory by a set of core tenets each asserting that one set of independent variables affects one dependent variable, moreover every variable can have several operational definitions. Data consist of a selected sample of scientific articles from the empirical literature on the theory under investigation. Our “tenet map” features a number of complexities more than the original version. First the links are two-layer: first-layer links connect variables which are related in the test of the theory at issue;second-layer links represent connections which are found statistically significant. Besides, either layer matrix of link-formation probabilities is block-symmetric. In addition to a form of censoring which resembles the Hui et al. pruning step, observed maps are subject to a further censoring related to second-layer links. Still, we perform a full Bayesian analysis instead of adopting the empirical Bayes approach. Lastly, we develop a three-stage model which accounts for dependence either of data or of parameters. The investigation of the empirical support and consensus degree of new economic theories of the firm motivated the proposed methodology. In this paper, the Transaction Cost Economics view is tested by a tenet map analysis. Both the two-stage and the multilevel models identify the same tenets as the most corroborated by empirical evidence though the latter provides a more comprehensive and complex insight of relationships between constructs.展开更多
Determining how animals respond to resource availability across spatial and temporal extents is crucial to understand ecological processes underpinning habitat selection.Here,we used a multi-scale approach to study th...Determining how animals respond to resource availability across spatial and temporal extents is crucial to understand ecological processes underpinning habitat selection.Here,we used a multi-scale approach to study the year-round habitat selection of the Crested Tit(Lophophanes cristatus)in a semi-natural lowland woodland of northern Italy,analysing different habitat features at each scale.We performed Crested Tit censuses at three different spatial scales.At the macrohabitat scale,we used geolocalized observations of individuals to compute Manly's habitat selection index,based on a detailed land-use map of the study area.At the microhabitat scale,the trees features were compared between presence and absence locations.At the foraging habitat scale,individual foraging birds and their specific position on trees were recorded using focal animal sampling.Censuses were performed during both the breeding(March to May)and wintering(December to January)seasons.At the macrohabitat scale,the Crested Tits significantly selected pure and mixed pine forests and avoided woods of alien plant species,farmlands and urban areas.At the microhabitat scale,old pine woods with dense cover were selected,with no significant difference in the features of tree selection between the two phenological phases.At the foraging habitat scale,the species was observed spending more time foraging in the canopies than in the understorey,using mostly the portion of Scots Pine(Pinus sylvestris)canopies closer to the trunk in winter,while during the breeding period,the whole canopy was visited.Overall,breeding and wintering habitats largely overlapped in the Crested Tit.Based on our findings,lowland Crested Tits can be well defined as true habitat specialists:they are strictly related to some specific coniferous woodland features.Noteworthily,compared to other tit species,which normally show generalist habits during winter,the Crested Tit behaves as a habitat specialist also out of the breeding season.Our study stressed the importance of considering multi-scale(both spatial and phenological)habitat selection in birds.展开更多
It is shown that any multicriteria problem can be represented by a hierarchical system of criteria. Individual properties of the object (alternative) are evaluated at the bottom level of the system, using a criteria v...It is shown that any multicriteria problem can be represented by a hierarchical system of criteria. Individual properties of the object (alternative) are evaluated at the bottom level of the system, using a criteria vector. A composition mechanism is used to evaluate the object as a whole at the top level. The problem is solved by the method of nested scalar convolutions of vector-valued criteria. The methodology of the problem solving is based on the complementarity principle by N. Bohr and the theorem of incompleteness by K. G?del. An example is presented that helps the reader digest some of the intricacies in the methodology.展开更多
A framework for accelerating modern long-running astrophysical simulations is presented, which is based on a hierarchical architecture where computational steering in the high-resolution run is performed under the gui...A framework for accelerating modern long-running astrophysical simulations is presented, which is based on a hierarchical architecture where computational steering in the high-resolution run is performed under the guide of knowledge obtained in the gradually refined ensemble analyses. Several visualization schemes for facilitating ensemble management, error analysis, parameter grouping and tuning are also integrated owing to the pluggable modular design. The proposed approach is prototyped based on the Flash code, and it can be extended by introducing userdefined visualization for specific requirements. Two real-world simulations, i.e., stellar wind and supernova remnant, are carried out to verify the proposed approach.展开更多
文摘Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.
基金National Natural Science Foundation of China(No.61261029)
文摘Watershed segmentation is sensitive to noises and irregular details within the image,which frequently leads to a serious over-segmentation Linear filtering before watershed segmentation can reduce over-segmentation to some extent,however,it often causes the position offset of object contours.For the purpose of reducing over-segmentation to preserve the location of object contours,the watershed segmentation based on the hierarchical multi-scale modification of morphological gradient is proposed.Firstly,multi-scale morphological filtering was employed to smooth the original image.Then,the gradient image was divided into multi-levels by the volume of three-dimension topographic relief,where the lower gradient layers were further modifiedby morphological closing with larger-sized structuring-elements,and the higher layers with the smaller one.In this way,most local minimums caused by irregular details and noises can be removed,while region contour positions corresponding to the target area were largely preserved.Finally,morphological watershed algorithm was employed to implement segmentation on the modified gradient image.The experimental results show that the proposed method can greatly reduce the over-segmentation of the watershed and avoid the position offset of the object contours.
基金This work was supported by the National Natural Science Foundation of China(Nos.62073322 and 61633020)the CIE-Tencent Robotics X Rhino-Bird Focused Research Program(No.2022-07)the Beijing Natural Science Foundation(No.2022MQ05).
文摘Grasp detection plays a critical role for robot manipulation.Mainstream pixel-wise grasp detection networks with encoder-decoder structure receive much attention due to good accuracy and efficiency.However,they usually transmit the high-level feature in the encoder to the decoder,and low-level features are neglected.It is noted that low-level features contain abundant detail information,and how to fully exploit low-level features remains unsolved.Meanwhile,the channel information in high-level feature is also not well mined.Inevitably,the performance of grasp detection is degraded.To solve these problems,we propose a grasp detection network with hierarchical multi-scale feature fusion and inverted shuffle residual.Both low-level and high-level features in the encoder are firstly fused by the designed skip connections with attention module,and the fused information is then propagated to corresponding layers of the decoder for in-depth feature fusion.Such a hierarchical fusion guarantees the quality of grasp prediction.Furthermore,an inverted shuffle residual module is created,where the high-level feature from encoder is split in channel and the resultant split features are processed in their respective branches.By such differentiation processing,more high-dimensional channel information is kept,which enhances the representation ability of the network.Besides,an information enhancement module is added before the encoder to reinforce input information.The proposed method attains 98.9%and 97.8%in image-wise and object-wise accuracy on the Cornell grasping dataset,respectively,and the experimental results verify the effectiveness of the method.
基金The National Natural Science Foundation of China(No.61603091)。
文摘In order to improve the detection accuracy of small objects,a neighborhood fusion-based hierarchical parallel feature pyramid network(NFPN)is proposed.Unlike the layer-by-layer structure adopted in the feature pyramid network(FPN)and deconvolutional single shot detector(DSSD),where the bottom layer of the feature pyramid network relies on the top layer,NFPN builds the feature pyramid network with no connections between the upper and lower layers.That is,it only fuses shallow features on similar scales.NFPN is highly portable and can be embedded in many models to further boost performance.Extensive experiments on PASCAL VOC 2007,2012,and COCO datasets demonstrate that the NFPN-based SSD without intricate tricks can exceed the DSSD model in terms of detection accuracy and inference speed,especially for small objects,e.g.,4%to 5%higher mAP(mean average precision)than SSD,and 2%to 3%higher mAP than DSSD.On VOC 2007 test set,the NFPN-based SSD with 300×300 input reaches 79.4%mAP at 34.6 frame/s,and the mAP can raise to 82.9%after using the multi-scale testing strategy.
文摘A concept map is a diagram depicting relationships among concepts which is used as a knowledge representation tool in many knowledge domains. In this paper, we build on the modeling framework of Hui et al. (2008) in order to develop a concept map suitable for testing the empirical evidence of theories. We identify a theory by a set of core tenets each asserting that one set of independent variables affects one dependent variable, moreover every variable can have several operational definitions. Data consist of a selected sample of scientific articles from the empirical literature on the theory under investigation. Our “tenet map” features a number of complexities more than the original version. First the links are two-layer: first-layer links connect variables which are related in the test of the theory at issue;second-layer links represent connections which are found statistically significant. Besides, either layer matrix of link-formation probabilities is block-symmetric. In addition to a form of censoring which resembles the Hui et al. pruning step, observed maps are subject to a further censoring related to second-layer links. Still, we perform a full Bayesian analysis instead of adopting the empirical Bayes approach. Lastly, we develop a three-stage model which accounts for dependence either of data or of parameters. The investigation of the empirical support and consensus degree of new economic theories of the firm motivated the proposed methodology. In this paper, the Transaction Cost Economics view is tested by a tenet map analysis. Both the two-stage and the multilevel models identify the same tenets as the most corroborated by empirical evidence though the latter provides a more comprehensive and complex insight of relationships between constructs.
文摘Determining how animals respond to resource availability across spatial and temporal extents is crucial to understand ecological processes underpinning habitat selection.Here,we used a multi-scale approach to study the year-round habitat selection of the Crested Tit(Lophophanes cristatus)in a semi-natural lowland woodland of northern Italy,analysing different habitat features at each scale.We performed Crested Tit censuses at three different spatial scales.At the macrohabitat scale,we used geolocalized observations of individuals to compute Manly's habitat selection index,based on a detailed land-use map of the study area.At the microhabitat scale,the trees features were compared between presence and absence locations.At the foraging habitat scale,individual foraging birds and their specific position on trees were recorded using focal animal sampling.Censuses were performed during both the breeding(March to May)and wintering(December to January)seasons.At the macrohabitat scale,the Crested Tits significantly selected pure and mixed pine forests and avoided woods of alien plant species,farmlands and urban areas.At the microhabitat scale,old pine woods with dense cover were selected,with no significant difference in the features of tree selection between the two phenological phases.At the foraging habitat scale,the species was observed spending more time foraging in the canopies than in the understorey,using mostly the portion of Scots Pine(Pinus sylvestris)canopies closer to the trunk in winter,while during the breeding period,the whole canopy was visited.Overall,breeding and wintering habitats largely overlapped in the Crested Tit.Based on our findings,lowland Crested Tits can be well defined as true habitat specialists:they are strictly related to some specific coniferous woodland features.Noteworthily,compared to other tit species,which normally show generalist habits during winter,the Crested Tit behaves as a habitat specialist also out of the breeding season.Our study stressed the importance of considering multi-scale(both spatial and phenological)habitat selection in birds.
文摘It is shown that any multicriteria problem can be represented by a hierarchical system of criteria. Individual properties of the object (alternative) are evaluated at the bottom level of the system, using a criteria vector. A composition mechanism is used to evaluate the object as a whole at the top level. The problem is solved by the method of nested scalar convolutions of vector-valued criteria. The methodology of the problem solving is based on the complementarity principle by N. Bohr and the theorem of incompleteness by K. G?del. An example is presented that helps the reader digest some of the intricacies in the methodology.
基金Supported by the National Natural Science Foundation of China(No.U1231108)
文摘A framework for accelerating modern long-running astrophysical simulations is presented, which is based on a hierarchical architecture where computational steering in the high-resolution run is performed under the guide of knowledge obtained in the gradually refined ensemble analyses. Several visualization schemes for facilitating ensemble management, error analysis, parameter grouping and tuning are also integrated owing to the pluggable modular design. The proposed approach is prototyped based on the Flash code, and it can be extended by introducing userdefined visualization for specific requirements. Two real-world simulations, i.e., stellar wind and supernova remnant, are carried out to verify the proposed approach.
