AMPNet: Average- and Max-Pool Networks for Salient Object Detection

Sun, Luna; Chen, Zhenxue; Wu, Qi; Zhao, Huijing; He, Weikai; Yan, Xinghe

doi:10.1109/tcsvt.2021.3054471

Cited by 42 publications

(8 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As discussed in the Methodology section, the YOLO architecture, consists of one Input Layer, ten Convolutional layers and six Pooling Layers performing Maxpooling operations [ 63 ]. The feature map obtained as output from the final convolutional layer is used as input for the YOLO algorithm.…”

Section: Resultsmentioning

confidence: 99%

A deep learning based multimodal interaction system for bed ridden and immobile hospital admitted patients: design, development and evaluation

Islam¹,

Aadeeb²,

Munna³

et al. 2022

BMC Health Serv Res

View full text Add to dashboard Cite

Background Hospital cabins are a part and parcel of the healthcare system. Most patients admitted in hospital cabins reside in bedridden and immobile conditions. Though different kinds of systems exist to aid such patients, most of them focus on specific tasks like calling for emergencies, monitoring patient health, etc. while the patients’ limitations are ignored. Though some patient interaction systems have been developed, only singular options like touch, hand gesture or voice based interaction were provided which may not be usable for bedridden and immobile patients. Methods At first, we reviewed the existing literature to explore the prevailing healthcare and interaction systems developed for bedridden and immobile patients. Then, a requirements elicitation study was conducted through semi-structured interviews. Afterwards, design goals were established to address the requirements. Based on these goals and by using computer vision and deep learning technologies, a hospital cabin control system having multimodal interactions facility was designed and developed for hospital admitted, bedridden and immobile patients. Finally, the system was evaluated through an experiment replicated with 12 hospital admitted patients to measure its effectiveness, usability and efficiency. Results As outcomes, firstly, a set of user-requirements were identified for hospital admitted patients and healthcare practitioners. Secondly, a hospital cabin control system was designed and developed that supports multimodal interactions for bedridden and immobile hospital admitted patients which includes (a) Hand gesture based interaction for moving a cursor with hand and showing hand gesture for clicking, (b) Nose teeth based interaction where nose is used for moving a cursor and teeth is used for clicking and (c) Voice based interaction for executing tasks using specific voice commands. Finally, the evaluation results showed that the system is efficient, effective and usable to the focused users with 100% success rate, reasonable number of attempts and task completion time. Conclusion In the resultant system, Deep Learning has been incorporated to facilitate multimodal interaction for enhancing accessibility. Thus, the developed system along with its evaluation results and the identified requirements provides a promising solution for the prevailing crisis in the healthcare sector. Trial Registration Not Applicable.

show abstract

Section: Resultsmentioning

confidence: 99%

A deep learning based multimodal interaction system for bed ridden and immobile hospital admitted patients: design, development and evaluation

Islam¹,

Aadeeb²,

Munna³

et al. 2022

BMC Health Serv Res

View full text Add to dashboard Cite

show abstract

“…In order to demonstrate the superior performance and fairness of the proposed method, this section conducted comparisons on four common salient object datasets, including DUTS-TE, DUT-OMRON, HKU-IS, and ECSSD, as well as 11 mainstream methods based on VGG-16 or ResNet-50 in the past three years, including AFNet [ 19 ], PAGE [ 20 ], MLMSNet [ 21 ], CPD [ 22 ], GateNet [ 23 ], ITSD [ 24 ], AMPNet [ 25 ], EGNet [ 6 ], BANet [ 26 ], BASNet [ 9 ], and DNA [ 27 ]. The evaluation metrics used in all comparison experiments included

, MAE, and the maximum F-measure.…”

Section: Experiments and Analysismentioning

confidence: 99%

A Salient Object Detection Method Based on Boundary Enhancement

Wen

Wang

Zou

et al. 2023

Sensors

View full text Add to dashboard Cite

Visual saliency refers to the human’s ability to quickly focus on important parts of their visual field, which is a crucial aspect of image processing, particularly in fields like medical imaging and robotics. Understanding and simulating this mechanism is crucial for solving complex visual problems. In this paper, we propose a salient object detection method based on boundary enhancement, which is applicable to both 2D and 3D sensors data. To address the problem of large-scale variation of salient objects, our method introduces a multi-level feature aggregation module that enhances the expressive ability of fixed-resolution features by utilizing adjacent features to complement each other. Additionally, we propose a multi-scale information extraction module to capture local contextual information at different scales for back-propagated level-by-level features, which allows for better measurement of the composition of the feature map after back-fusion. To tackle the low confidence issue of boundary pixels, we also introduce a boundary extraction module to extract the boundary information of salient regions. This information is then fused with salient target information to further refine the saliency prediction results. During the training process, our method uses a mixed loss function to constrain the model training from two levels: pixels and images. The experimental results demonstrate that our salient target detection method based on boundary enhancement shows good detection effects on targets of different scales, multi-targets, linear targets, and targets in complex scenes. We compare our method with the best method in four conventional datasets and achieve an average improvement of 6.2% on the mean absolute error (MAE) indicators. Overall, our approach shows promise for improving the accuracy and efficiency of salient object detection in a variety of settings, including those involving 2D/3D semantic analysis and reconstruction/inpainting of image/video/point cloud data.

show abstract

“…Especially in the field of remote sensing, more stringent requirements are put forward for these two points. The deep unfolding method has been widely studied [40]- [42]. It combines traditional optimization-based and learning-based methods.…”

Section: Introductionmentioning

confidence: 99%

Multisource Information Fusion Network for Optical Remote Sensing Image Super-Resolution

Shi

Gao

Chen

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

The super-resolution algorithms based on deep learning can effectively increase optical remote sensing image (ORSI) details for further analysis tasks. Deep unfolding methods have been studied in recent years to bridge the gap between optimization-based and learning-based methods, which can take good advantage of the prior knowledge. However, these unfolding methods usually ignore the utilization of intermediate network features between different iteration stages, thereby limiting the performance of super-resolution results. We propose a multisource information fusion network (MSFNet) for ORSI superresolution to address this problem. We mainly consider three strategies to enhance the image super-resolution performance, including feature extraction strategy, information fusion strategy, and the structure of the unfolding network. Firstly, image information of various scales is helpful for mining potential features of images for image super-resolution. Therefore, we introduce multi-scale implicit constraints to the objective function. Secondly, we unfold the optimization process into a neural network by alternating direction method of multipliers (ADMM). The network framework follows the ADMM method's iteration process. This unfolding strategy can effectively utilize the prior information for image reconstruction. Thirdly, we propose a rowcolumn decoupling Transformer module based on this unfolding framework for feature fusion. Specifically, the row Transformer block completes the feature fusion of various scales, and the column Transformer block completes the feature fusion of various channels. The fused features are transmitted to the next iteration stage for feature enhancement. And this fusion strategy enables the network to focus on the global information of the features. We perform experiments on three remote sensing image datasets to fully demonstrate the algorithm's effectiveness. Experiment results show that the proposed algorithm can achieve better image reconstruction performance.

show abstract

AMPNet: Average- and Max-Pool Networks for Salient Object Detection

Cited by 42 publications

References 50 publications

A deep learning based multimodal interaction system for bed ridden and immobile hospital admitted patients: design, development and evaluation

A deep learning based multimodal interaction system for bed ridden and immobile hospital admitted patients: design, development and evaluation

A Salient Object Detection Method Based on Boundary Enhancement

Multisource Information Fusion Network for Optical Remote Sensing Image Super-Resolution

Contact Info

Product

Resources

About