Lightweight Attention Pyramid Network for Object Detection and Instance Segmentation

Zhang, Jiwei; Yan, Yanyu; Cheng, Zelei; Wang, Wendong

doi:10.3390/app10030883

Cited by 19 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although there is a specific false-positive rate, the research in Ref. [37] shows that the calculation method of this false-positive rate is as follows:…”

Section: Verifiability Of Search Resultsmentioning

confidence: 99%

See 1 more Smart Citation

BFR‐SE: A Blockchain‐Based Fair and Reliable Searchable Encryption Scheme for IoT with Fine‐Grained Access Control in Cloud Environment

Gao

et al. 2021

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Due to capacity limitations, large amounts of data generated by IoT devices are often stored on cloud servers. These data are usually encrypted to prevent the disclosure, which significantly affects the availability of this data. Searchable encryption (SE) allows a party to store his data created by his IoT devices or mobile in encryption on the cloud server to protect his privacy while retaining his ability to search for data. However, the general SE techniques are all pay-then-use. The searchable encryption service providers (SESP) are considered curious but honest, making it unfair and unreliable. To address these problems, we combined ciphertext-policy attribute-based encryption, Bloom filter, and blockchain to propose a blockchain-based fair and reliable searchable encryption scheme (BFR-SE) in this paper. In BFR-SE, we constructed an attribute-based searchable encryption model that can provide fine-grained access control. The data owner stores the indices on SESP and stores some additional auxiliary information on the blockchain. After a data user initiates a request, SESP must return the correct and integral search results before the deadline. Otherwise, the data user can send an arbitration request, and the blockchain will make a ruling. The blockchain will only perform arbitrations based on auxiliary information when disputes arise, saving the computing resources on-chain. We analyzed the security and privacy of BFR-SE and simulated our scheme on the EOS blockchain, which proves that BFR-SE is feasible. Meanwhile, we provided a thorough analysis of storage and computing overhead, proving that BFR-SE is practical and has good performance.

show abstract

“…Although there is a specific false-positive rate, the research in Ref. [37] shows that the calculation method of this false-positive rate is as follows:…”

Section: Verifiability Of Search Resultsmentioning

confidence: 99%

“…Simultaneously, due to an attribute-based encryption algorithm, topics such as the revocation or update of permission are also one of the directions that need to be studied in the future. We will continue to refine our approach in conjunction with some other research [37,[54][55][56][57].…”

Section: Discussionmentioning

confidence: 99%

BFR‐SE: A Blockchain‐Based Fair and Reliable Searchable Encryption Scheme for IoT with Fine‐Grained Access Control in Cloud Environment

Gao

et al. 2021

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

show abstract

“…To improve the mask prediction accuracy, Zhang et al [43] designed a mask encoding pyramid network. To reduce the model calculation parameters, Zhang et al [44] proposed a lightweight FPN composed of attention transfer module and feature attention block. Kirillov et al [45] constructed a shared feature pyramid, which realized the interaction between different feature information by weight sharing.…”

Section: Feature Pyramid Network For Instance Segmentationmentioning

confidence: 99%

RPFNet: Recurrent Pyramid Frequency Feature Fusion Network for Instance Segmentation in Side-Scan Sonar Images

Wang

Zhang

et al. 2024

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

View full text Add to dashboard Cite

Side-scan sonar (SSS) is an essential acoustic sensor device for obtaining underwater information. The instance segmentation of sonar images can effectively locate and detect underwater objects. Although various CNN-based frameworks have achieved promising results in natural image instance segmentation, the noise interference, highlight shadow, and blurred edge in sonar images bring more significant challenges for sonar image instance segmentation. To solve these problems, we propose a novel recurrent pyramid frequency feature fusion network (RPFNet), which mainly consists of the pyramid frequency feature fusion network (PFN), recurrent residual attention mechanism (RRAM), mask prediction module, and semantic segmentation module. By enhancing and fusing different frequency features of SSS images, PFN can effectively extract fine-grained features and reduce background information interference. The RRAM uses residual structure and attention mechanism to enhance the correlation of different frequency features and improve nonlinear feature representation ability. The mask prediction and semantic segmentation modules are used to discriminate object instance categories and generate corresponding semantic segmentation results. We extensively evaluate the proposed framework on the real-scenes sonar image dataset. The evaluation results demonstrate that the proposed method outperforms the state-ofthe-art methods in various evaluation metrics. Code is available at https://github.com/darkseid-arch/SonarRPFNet.

show abstract

“…In target segmentation research, the articles [18][19][20][21][22][23] use edge information as a guide to sharpen the target, and edge information is mostly integrated from the contextual semantics to obtain the localization information to assist in the fusion of the synthesized high and low level feature information and thus achieve segmentation, whereas the article [23] obtains the edge semantics by using mask-guided pyramid networks; in target detection research, the articles [24,25] used the encoder part of different scales to progressively fuse features with the target significant edge extraction network to form a U-shaped structure to merge the object features and enhance the edges to cope with the rough boundaries of the object; However, most of the current research in the direction of target detection is aimed at the edges of significant targets, and it is still worthwhile to further improve and explore the structure of the guidance network for infrared ship images with low contrast and blurred contour boundaries.…”

Section: A Edge-guided Schemementioning

confidence: 99%

EGISD-YOLO: Edge Guidance Network for Infrared Ship Target Detection

Zhan,

Zhang,

Guo

et al. 2024

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

View full text Add to dashboard Cite

In the marine field, infrared detection technology is of great significance for timely localization and detection of ships in security missions. However, since infrared ship targets are often in the environmental conditions of small pixel occupancy, low contrast and complex background, it poses a great challenge for multi-ship detection, classification and localization tasks. Therefore, to address these problems we propose an edge informationguided infrared ship target detection network (EGISD-YOLO), in which a dense-csp structure is designed to improve the csp module of YOLO to increase the reusability of the backbone feature information, and in addition, to address the noise and interference generated by the image in the complex background, a deconvolutional channel attention module (DCA) is designed to link the contextual language to the image. DCA), which relates the contextual semantics to obtain the local information of the target. Crucially, we propose an edge-guided structure that takes the edge information of low-level features as a cue to fuse with deep-level features to enrich the target contour and thus improve the target localization ability, so that the network still possesses robustness under low-contrast conditions, and finally, we add a small-size prediction head at the end of the network to further increase the detection ability of weak targets. The proposed EGISD-YOLO is experimentally demonstrated to have better detection performance for infrared ship targets.

show abstract

Lightweight Attention Pyramid Network for Object Detection and Instance Segmentation

Cited by 19 publications

References 40 publications

BFR‐SE: A Blockchain‐Based Fair and Reliable Searchable Encryption Scheme for IoT with Fine‐Grained Access Control in Cloud Environment

BFR‐SE: A Blockchain‐Based Fair and Reliable Searchable Encryption Scheme for IoT with Fine‐Grained Access Control in Cloud Environment

RPFNet: Recurrent Pyramid Frequency Feature Fusion Network for Instance Segmentation in Side-Scan Sonar Images

EGISD-YOLO: Edge Guidance Network for Infrared Ship Target Detection

Contact Info

Product

Resources

About