Adaptive Spatiotemporal Feature Extraction and Dynamic Combining Methods for Selective Visual Attention System

Cheoi, Kyung Joo; Kim, Mi-hye

doi:10.1007/s11277-017-5043-0

Cited by 2 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed model was developed to effectively detect a distant target using the topdown information by expanding the bottom-up visual attention model proposed in [27]. To detect targets in the proposed model, the training process is required before searching for targets.…”

Section: The Methodologymentioning

confidence: 99%

“…The basic bottom-up process of extracting the early visual features and weighting them together to produce the saliency map proceeds in the same way as in the model of [27].…”

Section: Feature Extraction and Saliency Map Generationmentioning

confidence: 99%

“…Third, three form-orientation feature maps (F 21 for R/G color, F 22 for B/Y color, F 23 for intensity) are generated by extracting form-orientation features (F 21n , F 22n , F 23n , n=1~8) that have eight orientations from the extracted two color features (F 11 , F 12 ) and intensity feature (F 13 ) respectively. Center-surround computations with 8 orientations (0π/8, 1 π /8, … 7π/8) [27] are used as form-orientation filter.and combining them using an improved nonlinear combining method.…”

Section: Feature Extraction and Saliency Map Generationmentioning

confidence: 99%

“…In addition, a model using depth information was proposed to apply to the 3D image [23][24][25][26]. The main differences between the most successful computational models proposed so far are in methods of extracting features and generating the saliency map [27]. The word 'Saliency' means the intensity of a feature of an input image and can be described as a difference between a pixel and its surrounding neighborhood [1].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A novel distant target region detection method using hybrid saliency-based attention model under complex textures

Cheoi

2021

COMSIS J

View full text Add to dashboard Cite

In this paper, a hybrid visual attention model to effectively detect a distant target is proposed. The model employs the human visual attention mechanism and consists of two models, the training model, and the detection model. In the training model, some of the features are selected to train in the process of extracting and combining the early visual features from the training image of the target by bottom-up manner, and these features are trained and accumulated as trained data. When the image containing the target is input into the detection model, a task of selectively promoting only features of the target using pre-trained data is performed. As a result, the desired target is detected through the saliency map created as a result of the feature combination. The model has been tested on various images, and the experimental results demonstrate that the proposed model detected the target more accurately and faster than other previous models.

show abstract

Section: The Methodologymentioning

confidence: 99%

“…The basic bottom-up process of extracting the early visual features and weighting them together to produce the saliency map proceeds in the same way as in the model of [27].…”

Section: Feature Extraction and Saliency Map Generationmentioning

confidence: 99%

Section: Feature Extraction and Saliency Map Generationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A novel distant target region detection method using hybrid saliency-based attention model under complex textures

Cheoi

2021

COMSIS J

View full text Add to dashboard Cite

show abstract

“…It consists of a report, the classification and extraction of figures and tables, and key sentences for which the meaningful information among unstructured data is used. The paper by Cheoi et al [12] presents a selective visual-attention system using an adaptive spatiotemporal feature extraction and a dynamic combining method. Here, the user can detect the regions of interest (ROIs) effectively by adaptively selecting features and spatial saliencies according to experimental images.…”

mentioning

confidence: 99%

Knowledge-Based Communication System

Park

2018

Wireless Pers Commun

View full text Add to dashboard Cite

With the evolvement of wired and wireless communication technologies and the rapid development of heterogeneous big-data integration, the internet of things (IoT), digital transformation, and distributed storage processes, information communication technology (ICT) has gradually become complicated and intelligent. Knowledge-based communication is used as an evolutionary platform to interact with users through wireless personalized communication; this communication form's capability to connect with knowledge-based communication means that general users are able to access the platform easily and can receive a variety of services.The paper by Thirunavukkarasu et al.[1] presents a nonintrusive forensic-detection method using a discrete stationary wavelet transform (SWT) and multidimensional-scaling (MDS) feature descriptors. It is used as a reduced feature set for detecting the category of copy-move image tampering that automatically identifies multiple tampered regions and distorted tampered regions using the knowledge of the images. The paper by Lim et al. [2] presents the design of a wearable radio-frequency identification antenna using a redesigned antenna and the employment of spiral structures. This is used in the development of the wearable radio-frequency identification antenna for cuffs or shirt collars based on an analysis and the dimensions of the antenna geometry.The paper by Jeong et al.[3] presents a hierarchical biometric-information management scheme through a heterogeneous-device synchronization. This is the study of an analytic hierarchy process (AHP) that uses an information-management scheme for the analysis of medical-device correlations for the purpose of treatment. In addition, it performs not only hospital administrative work but also disease analysis and treatment. The paper by Jin et al.[4] introduces influencing factors, a structural connotation, and measuring dimensions

show abstract

Adaptive Spatiotemporal Feature Extraction and Dynamic Combining Methods for Selective Visual Attention System

Cited by 2 publications

References 15 publications

A novel distant target region detection method using hybrid saliency-based attention model under complex textures

A novel distant target region detection method using hybrid saliency-based attention model under complex textures

Knowledge-Based Communication System

Contact Info

Product

Resources

About