Graphical components information extraction is a crucial step in the chart recognition and understanding process. However, existing methods of information extraction from chart images either are type-dependent or rely...
详细信息
the most distinctive trait in structural patternrecognition in graph domain is the ability to deal withthe organization and relations between the constituent entities of the pattern. Even if this can be convenient a...
详细信息
ISBN:
(纸本)9789897584756
the most distinctive trait in structural patternrecognition in graph domain is the ability to deal withthe organization and relations between the constituent entities of the pattern. Even if this can be convenient and/or necessary in many contexts, most of the state-of the art classification techniques can not be deployed directly in the graph domain without first embedding graph patterns towards a metric space. Granular computing is a powerful information processing paradigm that can be employed in order to drive the synthesis of automatic embedding spaces from structured domains. In this paper we investigate several classification techniques starting from Granular computing-based embedding procedures and provide a thorough overview in terms of model complexity, embedding space complexity and performances on several open-access datasets for graph classification. We witness that certain classification techniques perform poorly both from the point of view of complexity and learning performances as the case of non-linear SVM, suggesting that high dimensionality of the synthesized embedding space can negatively affect the effectiveness of these approaches. On the other hand, linear support vector machines, neuro-fuzzy networks and nearest neighbour classifiers have comparable performances in terms of accuracy, with second being the most competitive in terms of structural complexity and the latter being the most competitive in terms of embedding space dimensionality.
this paper presents our work on automatically locating charts from document pages, which is an important stage in our chart image recognition and understanding system currently being developed. To achieve this, there ...
详细信息
the necessity and trends of developing the methods of the patternrecognitiontheory for solving nonstandard tasks of train sorting at the marshalling yard have been substantiated. the examples of similar tasks from o...
详细信息
Non-stationary data distributions are a challenge in activity recognition from body worn motion sensors. Classifier models have to be adapted online to maintain a high recognition performance. Typical approaches for o...
详细信息
In order to improve the performance of face recognition with single sample effectively, a face recognition method based on multiple features and twice classification is proposed. For obtaining sufficient face informat...
详细信息
ISBN:
(纸本)9783642394799;9783642394782
In order to improve the performance of face recognition with single sample effectively, a face recognition method based on multiple features and twice classification is proposed. For obtaining sufficient face information, facial multiple features combining differential excitation and Compound Local Binary pattern (CLBP) on the asymmetric region are extracted. Elastic Matching (EM) has better robustness for pose. However, the computation complexity of the method is rather high. Classifying twice strategy is proposed to short the time of data processing. Experimental results on ORL database and FERET database show that the method is effective in getting better recognition rate and speed, also has a certain robustness to pose.
Local binary patterns (LBP) are considered as the most computation efficient and high-performance texture features. Among all variants of the LBPs, Median Robust Extended Local Binary pattern (MRELBP) [1] is considere...
详细信息
the proceedings contain 1408 papers. the topics discussed include: deep gait relative attribute using a signed quadratic contrastive loss;variational capsule encoder;the DeepHealth toolkit: a unified framework to boos...
ISBN:
(纸本)9781728188089
the proceedings contain 1408 papers. the topics discussed include: deep gait relative attribute using a signed quadratic contrastive loss;variational capsule encoder;the DeepHealth toolkit: a unified framework to boost biomedical applications;hierarchically aggregated residual transformation for single image super resolution;occlusion-tolerant and personalized 3D human pose estimation in RGB images;computing stable resultant-based minimal solvers by hiding a variable;semantic segmentation for pedestrian detection from motion in temporal domain;multi-label contrastive focal loss for pedestrian attribute recognition;DmifNet: 3D shape reconstruction based on dynamic multi–branch information fusion;visual localization for autonomous driving: mapping the accurate location in the city maze;Bayesian active learning for maximal information gain on model parameters;and cross-spectrum face recognition using subspace projection hashing.
this book constitutes the proceedings of the 9thinternationalconference on Algorithms and Discrete Applied Mathematics, CALDAM 2023, which was held in Gandhinagar, India, during February 9-11, 2023.
ISBN:
(数字)9783031252112
ISBN:
(纸本)9783031252105
this book constitutes the proceedings of the 9thinternationalconference on Algorithms and Discrete Applied Mathematics, CALDAM 2023, which was held in Gandhinagar, India, during February 9-11, 2023.
Kernel methods are widely used for document classification in diverse domains. Popular kernels such as bag-of-word kernels and tree kernels show satisfactory results in classifying documents such as articles, e-mails ...
详细信息
暂无评论