Breast cancer is a significant threat to the global population,affecting not only women but also a threat to the entire population.With recent advancements in digital pathology,Eosin and hematoxylin images provide enh...Breast cancer is a significant threat to the global population,affecting not only women but also a threat to the entire population.With recent advancements in digital pathology,Eosin and hematoxylin images provide enhanced clarity in examiningmicroscopic features of breast tissues based on their staining properties.Early cancer detection facilitates the quickening of the therapeutic process,thereby increasing survival rates.The analysis made by medical professionals,especially pathologists,is time-consuming and challenging,and there arises a need for automated breast cancer detection systems.The upcoming artificial intelligence platforms,especially deep learning models,play an important role in image diagnosis and prediction.Initially,the histopathology biopsy images are taken from standard data sources.Further,the gathered images are given as input to the Multi-Scale Dilated Vision Transformer,where the essential features are acquired.Subsequently,the features are subjected to the Bidirectional Long Short-Term Memory(Bi-LSTM)for classifying the breast cancer disorder.The efficacy of the model is evaluated using divergent metrics.When compared with other methods,the proposed work reveals that it offers impressive results for detection.展开更多
Ruth and Mary are two heroines in Eugene O'Neill's plays Beyond the Horizon, and Long Day's Journey into Night. They have some similarities: when they are young, they are beautiful, native and full of hope...Ruth and Mary are two heroines in Eugene O'Neill's plays Beyond the Horizon, and Long Day's Journey into Night. They have some similarities: when they are young, they are beautiful, native and full of hope towards the future life, but both make wrong choices; in the following years, both suffer a lot from these wrong choices, and feel regretful. This paper tries to explore these two tragic female images.展开更多
Diffusion tensor imaging plays an important role in the accurate diagnosis and prognosis of spinal cord diseases. However, because of technical limitations, the imaging sequences used in this technique cannot reveal t...Diffusion tensor imaging plays an important role in the accurate diagnosis and prognosis of spinal cord diseases. However, because of technical limitations, the imaging sequences used in this technique cannot reveal the fine structure of the spinal cord with precision. We used the readout segmentation of long variable echo-trains（RESOLVE） sequence in this cross-sectional study of 45 healthy volunteers aged 20 to 63 years. We found that the RESOLVE sequence significantly increased the resolution of the diffusion images and improved the median signal-to-noise ratio of the middle（C4–6） and lower（C7–T1） cervical segments to the level of the upper cervical segment. In addition, the values of fractional anisotropy and radial diffusivity were significantly higher in white matter than in gray matter. Our study verified that the RESOLVE sequence could improve resolution of diffusion tensor imaging in clinical applications and provide accurate baseline data for the diagnosis and treatment of cervical spinal cord diseases.展开更多
In view of the difficulties in traditional long-wave infrared imaging spectrometer which is hard to realize a high signal-to-noise ratio and miniaturization as well under the weak remote sensing signal,Offner convex g...In view of the difficulties in traditional long-wave infrared imaging spectrometer which is hard to realize a high signal-to-noise ratio and miniaturization as well under the weak remote sensing signal,Offner convex grating spectrometer and Dyson concave grating spectrometer,both having concentric structure,are designed and analyzed in the band of 8-12 μm. The diffraction angle expressions of the two spectrometers are obtained and the diffraction characteristics are acquired. Both of the spectrometers are designed in Zemax environment under different F-numbers and different grating constants with the same slit,spatial resolution,spectral resolution and detector. The results show that Dyson grating spectrometer possesses the advantages of higher throughput and smaller volume, and Offner grating spectrometer possesses the advantage of more accessible material and the absence of chromatic aberration. The differences between Dyson form and Offner form show that the former is a better choice in the long-wave infrared imaging spectrometer.展开更多
The successful launch and commissioning of the first geostationary meteorological satellite of Korea has the potential to enhance earth observation capability over the Asia Pacific region. Although the specifications ...The successful launch and commissioning of the first geostationary meteorological satellite of Korea has the potential to enhance earth observation capability over the Asia Pacific region. Although the specifications of the payload, the mete- orological imager （MI）, have been verified during both ground and in-orbit tests, there is the possibility of variation and/or degradation of data quality due to many different reasons, such as the accumulation of contaminants, the aging of instrument components, and unexpected external disturbance. Thus, for better utilization of MI data, it is imperative to continuously monitor and maintain the data quality. As a part of such activity, this study presents an inter-calibration, based on the Global Space-based Inter-Calibration System （GSICS）, between the MI data and the high quality hyperspectral data from the In- frared Atmospheric Sounding Interferometer （IASI） of the Metop-A satellite. Both sets of data, acquired for three years from April 2011 to March 2014, are processed to prepare the matchup dataset, which is spatially collocated, temporally concurrent, angularly coincident, and spectrally comparable. The results show that the MI data are stable within the specifications and show no significant degradation during the study period. However, the water vapor channel shows a rather large bias value of -0.77 K, with a root-mean-square difference （RMSD） of around 1.1 K, which is thought to be due to the shift in the spectral response function. The shortwave channel shows a maximum RMSD of around 1.39 K, mainly due to the coarse digitization at the lower temperature. The inter-comparison results are re-checked through a sensitivity analysis with different sets of threshold values used for the matchup dataset. Based on this, we confirm that the overall quality of the MI data meets the user requirements and maintains the expected performance, although the water vapor channel requires further investigation.展开更多
A novel content based image retrieval （CBIR） algorithmusing relevant feedback is presented. The proposed frameworkhas three major contributions： a novel feature descriptor calledcolor spectral histogram （CSH） to ...A novel content based image retrieval （CBIR） algorithmusing relevant feedback is presented. The proposed frameworkhas three major contributions： a novel feature descriptor calledcolor spectral histogram （CSH） to measure the similarity betweenimages; two-dimensional matrix based indexing approach proposedfor short-term learning （STL）; and long-term learning （LTL）.In general, image similarities are measured from feature representationwhich includes color quantization, texture, color, shapeand edges. However, CSH can describe the image feature onlywith the histogram. Typically the image retrieval process starts byfinding the similarity between the query image and the imagesin the database; the major computation involved here is that theselection of top ranking images requires a sorting algorithm to beemployed at least with the lower bound of O（n log n）. A 2D matrixbased indexing of images can enormously reduce the searchtime in STL. The same structure is used for LTL with an aim toreduce the amount of log to be maintained. The performance ofthe proposed framework is analyzed and compared with the existingapproaches, the quantified results indicates that the proposedfeature descriptor is more effectual than the existing feature descriptorsthat were originally developed for CBIR. In terms of STL,the proposed 2D matrix based indexing minimizes the computationeffort for retrieving similar images and for LTL, the proposed algorithmtakes minimum log information than the existing approaches.展开更多
针对SAR图像检测船舶任务中的目标小、近岸样本目标检测困难等问题,文章提出一种名为长短路特征融合网络(Long and Short path Feature Fusion Network,LSFF-Net)的船舶检测网络。该网络通过长短路特征融合模块有效协调了大目标与小目...针对SAR图像检测船舶任务中的目标小、近岸样本目标检测困难等问题,文章提出一种名为长短路特征融合网络(Long and Short path Feature Fusion Network,LSFF-Net)的船舶检测网络。该网络通过长短路特征融合模块有效协调了大目标与小目标检测,避免小目标特征信息的丢失。网络中应用结构重参数化结构提高了模块学习能力。为了满足多尺度目标检测,加入特征金字塔网络,融合多尺度特征。为了应对近岸样本目标检测,设计数据重分配算法,提高了对近岸样本目标的检测精度。实验结果表明:在公开数据集检测时,算法的平均精度(Average Precision,AP)达到97.50%,优于主流目标检测算法。该方法为提高SAR图像中小目标和近岸样本目标检测精度提供了新的实现方案。展开更多
针对图像描述方法中对图像文本信息的遗忘及利用不充分问题,提出了基于场景图感知的跨模态交互网络(SGC-Net)。首先,使用场景图作为图像的视觉特征并使用图卷积网络(GCN)进行特征融合,从而使图像的视觉特征和文本特征位于同一特征空间;...针对图像描述方法中对图像文本信息的遗忘及利用不充分问题,提出了基于场景图感知的跨模态交互网络(SGC-Net)。首先,使用场景图作为图像的视觉特征并使用图卷积网络(GCN)进行特征融合,从而使图像的视觉特征和文本特征位于同一特征空间;其次,保存模型生成的文本序列,并添加对应的位置信息作为图像的文本特征,以解决单层长短期记忆(LSTM)网络导致的文本特征丢失的问题;最后,使用自注意力机制提取出重要的图像信息和文本信息后并对它们进行融合,以解决对图像信息过分依赖以及对文本信息利用不足的问题。在Flickr30K和MSCOCO(MicroSoft Common Objects in COntext)数据集上进行实验的结果表明,与Sub-GC相比,SGC-Net在BLEU1(BiLingual Evaluation Understudy with 1-gram)、BLEU4(BiLingual Evaluation Understudy with 4-grams)、METEOR(Metric for Evaluation of Translation with Explicit ORdering)、ROUGE(Recall-Oriented Understudy for Gisting Evaluation)和SPICE(Semantic Propositional Image Caption Evaluation)指标上分别提升了1.1、0.9、0.3、0.7、0.4和0.3、0.1、0.3、0.5、0.6。可见,SGC-Net所使用的方法能够有效提升模型的图像描述性能及生成描述的流畅度。展开更多
Image paragraph generation aims to generate a long description composed of multiple sentences,which is different from traditional image captioning containing only one sentence.Most of previous methods are dedicated to...Image paragraph generation aims to generate a long description composed of multiple sentences,which is different from traditional image captioning containing only one sentence.Most of previous methods are dedicated to extracting rich features from image regions,and ignore modelling the visual relationships.In this paper,we propose a novel method to generate a paragraph by modelling visual relationships comprehensively.First,we parse an image into a scene graph,where each node represents a specific object and each edge denotes the relationship between two objects.Second,we enrich the object features by implicitly encoding visual relationships through a graph convolutional network(GCN).We further explore high-order relations between different relation features using another graph convolutional network.In addition,we obtain the linguistic features by projecting the predicted object labels and their relationships into a semantic embedding space.With these features,we present an attention-based topic generation network to select relevant features and produce a set of topic vectors,which are then utilized to generate multiple sentences.We evaluate the proposed method on the Stanford image-paragraph dataset which is currently the only available dataset for image paragraph generation,and our method achieves competitive performance in comparison with other state-of-the-art(SOTA)methods.展开更多
基金Deanship of Research and Graduate Studies at King Khalid University for funding this work through Small Group Research Project under Grant Number RGP1/261/45.
文摘Breast cancer is a significant threat to the global population,affecting not only women but also a threat to the entire population.With recent advancements in digital pathology,Eosin and hematoxylin images provide enhanced clarity in examiningmicroscopic features of breast tissues based on their staining properties.Early cancer detection facilitates the quickening of the therapeutic process,thereby increasing survival rates.The analysis made by medical professionals,especially pathologists,is time-consuming and challenging,and there arises a need for automated breast cancer detection systems.The upcoming artificial intelligence platforms,especially deep learning models,play an important role in image diagnosis and prediction.Initially,the histopathology biopsy images are taken from standard data sources.Further,the gathered images are given as input to the Multi-Scale Dilated Vision Transformer,where the essential features are acquired.Subsequently,the features are subjected to the Bidirectional Long Short-Term Memory(Bi-LSTM)for classifying the breast cancer disorder.The efficacy of the model is evaluated using divergent metrics.When compared with other methods,the proposed work reveals that it offers impressive results for detection.
文摘Ruth and Mary are two heroines in Eugene O'Neill's plays Beyond the Horizon, and Long Day's Journey into Night. They have some similarities: when they are young, they are beautiful, native and full of hope towards the future life, but both make wrong choices; in the following years, both suffer a lot from these wrong choices, and feel regretful. This paper tries to explore these two tragic female images.
文摘Diffusion tensor imaging plays an important role in the accurate diagnosis and prognosis of spinal cord diseases. However, because of technical limitations, the imaging sequences used in this technique cannot reveal the fine structure of the spinal cord with precision. We used the readout segmentation of long variable echo-trains（RESOLVE） sequence in this cross-sectional study of 45 healthy volunteers aged 20 to 63 years. We found that the RESOLVE sequence significantly increased the resolution of the diffusion images and improved the median signal-to-noise ratio of the middle（C4–6） and lower（C7–T1） cervical segments to the level of the upper cervical segment. In addition, the values of fractional anisotropy and radial diffusivity were significantly higher in white matter than in gray matter. Our study verified that the RESOLVE sequence could improve resolution of diffusion tensor imaging in clinical applications and provide accurate baseline data for the diagnosis and treatment of cervical spinal cord diseases.
基金Sponsored by the National High Technology Research and Development Program of China(863 Program)(Grant No.2013AA03A116)
文摘In view of the difficulties in traditional long-wave infrared imaging spectrometer which is hard to realize a high signal-to-noise ratio and miniaturization as well under the weak remote sensing signal,Offner convex grating spectrometer and Dyson concave grating spectrometer,both having concentric structure,are designed and analyzed in the band of 8-12 μm. The diffraction angle expressions of the two spectrometers are obtained and the diffraction characteristics are acquired. Both of the spectrometers are designed in Zemax environment under different F-numbers and different grating constants with the same slit,spatial resolution,spectral resolution and detector. The results show that Dyson grating spectrometer possesses the advantages of higher throughput and smaller volume, and Offner grating spectrometer possesses the advantage of more accessible material and the absence of chromatic aberration. The differences between Dyson form and Offner form show that the former is a better choice in the long-wave infrared imaging spectrometer.
基金supported by the project entitled "Development of Meteorological Satellite Operation and Application Technology" of the KMA/NMSC (Korea Meteorological Adminstration/National Meteorological Satellite Center)supported by the Eco Innovation Program of KEITI (Korea Environmental Industry & Technology Institute) (Grant No. 2013000160002)
文摘The successful launch and commissioning of the first geostationary meteorological satellite of Korea has the potential to enhance earth observation capability over the Asia Pacific region. Although the specifications of the payload, the mete- orological imager （MI）, have been verified during both ground and in-orbit tests, there is the possibility of variation and/or degradation of data quality due to many different reasons, such as the accumulation of contaminants, the aging of instrument components, and unexpected external disturbance. Thus, for better utilization of MI data, it is imperative to continuously monitor and maintain the data quality. As a part of such activity, this study presents an inter-calibration, based on the Global Space-based Inter-Calibration System （GSICS）, between the MI data and the high quality hyperspectral data from the In- frared Atmospheric Sounding Interferometer （IASI） of the Metop-A satellite. Both sets of data, acquired for three years from April 2011 to March 2014, are processed to prepare the matchup dataset, which is spatially collocated, temporally concurrent, angularly coincident, and spectrally comparable. The results show that the MI data are stable within the specifications and show no significant degradation during the study period. However, the water vapor channel shows a rather large bias value of -0.77 K, with a root-mean-square difference （RMSD） of around 1.1 K, which is thought to be due to the shift in the spectral response function. The shortwave channel shows a maximum RMSD of around 1.39 K, mainly due to the coarse digitization at the lower temperature. The inter-comparison results are re-checked through a sensitivity analysis with different sets of threshold values used for the matchup dataset. Based on this, we confirm that the overall quality of the MI data meets the user requirements and maintains the expected performance, although the water vapor channel requires further investigation.
文摘A novel content based image retrieval （CBIR） algorithmusing relevant feedback is presented. The proposed frameworkhas three major contributions： a novel feature descriptor calledcolor spectral histogram （CSH） to measure the similarity betweenimages; two-dimensional matrix based indexing approach proposedfor short-term learning （STL）; and long-term learning （LTL）.In general, image similarities are measured from feature representationwhich includes color quantization, texture, color, shapeand edges. However, CSH can describe the image feature onlywith the histogram. Typically the image retrieval process starts byfinding the similarity between the query image and the imagesin the database; the major computation involved here is that theselection of top ranking images requires a sorting algorithm to beemployed at least with the lower bound of O（n log n）. A 2D matrixbased indexing of images can enormously reduce the searchtime in STL. The same structure is used for LTL with an aim toreduce the amount of log to be maintained. The performance ofthe proposed framework is analyzed and compared with the existingapproaches, the quantified results indicates that the proposedfeature descriptor is more effectual than the existing feature descriptorsthat were originally developed for CBIR. In terms of STL,the proposed 2D matrix based indexing minimizes the computationeffort for retrieving similar images and for LTL, the proposed algorithmtakes minimum log information than the existing approaches.
文摘针对SAR图像检测船舶任务中的目标小、近岸样本目标检测困难等问题,文章提出一种名为长短路特征融合网络(Long and Short path Feature Fusion Network,LSFF-Net)的船舶检测网络。该网络通过长短路特征融合模块有效协调了大目标与小目标检测,避免小目标特征信息的丢失。网络中应用结构重参数化结构提高了模块学习能力。为了满足多尺度目标检测,加入特征金字塔网络,融合多尺度特征。为了应对近岸样本目标检测,设计数据重分配算法,提高了对近岸样本目标的检测精度。实验结果表明:在公开数据集检测时,算法的平均精度(Average Precision,AP)达到97.50%,优于主流目标检测算法。该方法为提高SAR图像中小目标和近岸样本目标检测精度提供了新的实现方案。
文摘针对图像描述方法中对图像文本信息的遗忘及利用不充分问题,提出了基于场景图感知的跨模态交互网络(SGC-Net)。首先,使用场景图作为图像的视觉特征并使用图卷积网络(GCN)进行特征融合,从而使图像的视觉特征和文本特征位于同一特征空间;其次,保存模型生成的文本序列,并添加对应的位置信息作为图像的文本特征,以解决单层长短期记忆(LSTM)网络导致的文本特征丢失的问题;最后,使用自注意力机制提取出重要的图像信息和文本信息后并对它们进行融合,以解决对图像信息过分依赖以及对文本信息利用不足的问题。在Flickr30K和MSCOCO(MicroSoft Common Objects in COntext)数据集上进行实验的结果表明,与Sub-GC相比,SGC-Net在BLEU1(BiLingual Evaluation Understudy with 1-gram)、BLEU4(BiLingual Evaluation Understudy with 4-grams)、METEOR(Metric for Evaluation of Translation with Explicit ORdering)、ROUGE(Recall-Oriented Understudy for Gisting Evaluation)和SPICE(Semantic Propositional Image Caption Evaluation)指标上分别提升了1.1、0.9、0.3、0.7、0.4和0.3、0.1、0.3、0.5、0.6。可见,SGC-Net所使用的方法能够有效提升模型的图像描述性能及生成描述的流畅度。
基金supported in part by National Natural Science Foundation of China(Nos.61721004,61976214,62076078 and 62176246).
文摘Image paragraph generation aims to generate a long description composed of multiple sentences,which is different from traditional image captioning containing only one sentence.Most of previous methods are dedicated to extracting rich features from image regions,and ignore modelling the visual relationships.In this paper,we propose a novel method to generate a paragraph by modelling visual relationships comprehensively.First,we parse an image into a scene graph,where each node represents a specific object and each edge denotes the relationship between two objects.Second,we enrich the object features by implicitly encoding visual relationships through a graph convolutional network(GCN).We further explore high-order relations between different relation features using another graph convolutional network.In addition,we obtain the linguistic features by projecting the predicted object labels and their relationships into a semantic embedding space.With these features,we present an attention-based topic generation network to select relevant features and produce a set of topic vectors,which are then utilized to generate multiple sentences.We evaluate the proposed method on the Stanford image-paragraph dataset which is currently the only available dataset for image paragraph generation,and our method achieves competitive performance in comparison with other state-of-the-art(SOTA)methods.
