A Deep Evaluator for Image Retargeting Quality by Geometrical and Contextual Interaction

Jiang, Bin; Yang, Jiachen; Meng, Qinggang; Li, Baihua; Lu, Wen

doi:10.1109/tcyb.2018.2864158

Cited by 32 publications

(8 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…9. In particular, sub-figure 9 (1) describes the recognition of the target; and 9 (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) show the reaching process towards the target. The joint angles of the robotic arm were adjusted step by step based on the position errors between the hand and target, and the whole process took 11 steps.…”

Section: B Reaching Movement Experimentsmentioning

confidence: 99%

“…11 shows another experiment process, with a different target position closer to the robotic arm. The rough reaching movement and the correction movement are shown in subfigures 11 (2) and 11 (3)(4)(5)(6)(7)(8), respectively. In this experiment, the grasper successfully grasped the ball, even though the target was closer to the robotic arm, which made the robotic arm easier to touch the ball and lead to a failure.…”

Section: B Reaching Movement Experimentsmentioning

confidence: 99%

See 1 more Smart Citation

Visual-Guided Robotic Object Grasping Using Dual Neural Network Controllers

Fang

Chao

Lin

et al. 2021

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

It has been a challenging task for a robotic arm to accurately reach and grasp objects, and much research attention has been attracted. This paper proposes a robotic hand-eye coordination system by simulating the human behavior pattern to achieve a fast and robust reaching ability. This is achieved by two neural-network-based controllers, including a rough reaching movement controller implemented by a pre-trained Radial Basis Function (RBF) for rough reaching movements, and a correction movement controller built from a specifically designed Brain Emotional Nesting Network (BENN) for smooth correction movements. In particular, the proposed BENN is designed with high nonlinear mapping ability, with its adaptive laws derived from the Lyapunov stability theorem; from this, the robust tracking performance and accordingly the stability of the proposed control system are guaranteed by the utilisation of the H ∞ control approach. The proposed BENN is validated and evaluated by a chaos synchronisation simulation, and the overall control system by object grasping tasks through a physical robotic arm in a real-world environment. The experimental results demonstrate the superiority of the proposed control system in reference to those with single neural networks.

show abstract

Section: B Reaching Movement Experimentsmentioning

confidence: 99%

Section: B Reaching Movement Experimentsmentioning

confidence: 99%

Visual-Guided Robotic Object Grasping Using Dual Neural Network Controllers

Fang

Chao

Lin

et al. 2021

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

show abstract

“…Traditional assessments are often designed to evaluate the quality of enhanced images, which has made great achievements. Among these traditional criterions, four metrics are selected here to assess finger-vein image quality: the gradient in spatial domain, the gray contrast, the information capacity, and the deep evaluator based on stacked auto encoder [32].…”

Section: B Performance Metrics For Image Qualitymentioning

confidence: 99%

“…DESAE is a image retargeting quality assessment, and it evaluates image quality based on both geometry and content features, which is sensitive to the foundational information in images [32]. It can be computed as,…”

Section: ) Deep Evaluator Based On Stacked Auto Encodermentioning

confidence: 99%

Finger-Vein Image Enhancement Based on Pulse Coupled Neural Network

Lei

Chen

2019

IEEE Access

View full text Add to dashboard Cite

As a promising biometric technique, the finger-vein image recognition has seen a recent surge of interest. However, the collected finger-vein images are often degraded seriously because of various vein patterns, uneven illumination, and unsatisfied sensor conditions. This makes vein representations unreliable and inevitably impairs recognition accuracy. In this paper, a new model based on the pulse coupled neural network (PCNN) is proposed to enhance finger-vein image quality, and further to improve the reliability of image recognition. First, a simplified PCNN model for finger-vein image enhancement is proposed to reduce computational complexity. Then, a new parameter setting scheme is developed to automatically adjust the parameters in the present PCNN model, without involving any empirical correlations or training. Finally, extensive experiments are carried out, and the performance of the proposed PCNN model is evaluated by calculating four metrics, including the gradient in the spatial domain (GSG), the gray contrast (GC), the information capacity (IC), and the deep evaluator based on the stacked auto encoder (DESAE). The results have verified the capability and reliability of the presented PCNN model as well as the parameter setting scheme in finger-vein image enhancement. In addition, the enhanced finger-vein images by the present PCNN model have been shown to be able to improve the recognition accuracy significantly.INDEX TERMS Pulse coupled neural network, finger vein, automatic parameter setting, image enhancement.

show abstract

“…Recently, feature tracking based on feature detection and matching framework (DMF) [40], [41] have received much attention from the communities of compute vision [42]- [46] and computer graphics [47]. For example, Zhang et al [42] proposed a non-consecutive feature tracking method for SFM-based 3D reconstruction and simultaneous localization and mapping.…”

Section: A Feature Trackingmentioning

confidence: 99%

Parallel K Nearest Neighbor Matching for 3D Reconstruction

Cao

Xie

et al. 2019

IEEE Access

View full text Add to dashboard Cite

In recent years, a 3D reconstruction based on structure from motion (SFM) has attracted much attention from the communities of computer vision and graphics. It is well known that the speed and quality of SFM systems largely depend on the technique of feature tracking. If a big volume of image data is inputted for SFM, the speed of this SFM system would become very slow. And, this problem becomes severer for largescale scenes, which typically needs to capture several thousands of images to recover the point-cloud model of the scene. However, none of the existing methods fully addresses the problem of fast feature tracking. Brute force matching is capable of producing correspondences for small-scale scenes but often getting stuck in repeated features. Hashing matching can only deal with middle-scale scenes and is not capable of largescale scenes. In this paper, we propose a new feature tacking method working in a parallel manner rather than in a single thread scheme. Our method consists of steps of keypoint detection, descriptor computing, descriptor matching by parallel k-nearest neighbor (Parallel-KNN) search, and outlier rejecting. This method is able to rapidly match a big volume of keypoints and avoids to consume high computation time, then yielding a set of correct correspondences. We demonstrate and evaluate the proposed method on several challenging benchmark datasets, including those with highly repeated features, and compare to the state-ofthe-art methods. The experimental results indicate that our method outperforms the compared methods in both efficiency and effectiveness.

show abstract

A Deep Evaluator for Image Retargeting Quality by Geometrical and Contextual Interaction

Cited by 32 publications

References 62 publications

Visual-Guided Robotic Object Grasping Using Dual Neural Network Controllers

Visual-Guided Robotic Object Grasping Using Dual Neural Network Controllers

Finger-Vein Image Enhancement Based on Pulse Coupled Neural Network

Parallel K Nearest Neighbor Matching for 3D Reconstruction

Contact Info

Product

Resources

About