X-Ray Image Detection Algorithm for Prohibited Items Based on Feature Enhancement and Loss Optimization
Journal Title: Engineering and Technology Journal - Year 2024, Vol 9, Issue 08
Abstract
To address the issue of low detection accuracy of prohibited items in X-ray security images caused by varying orientations, different scales, and the intertwining of targets with backgrounds, we propose a novel X-ray image detection algorithm based on feature enhancement and loss optimization. The model is built upon the ConvNext network and incorporates a Directional Channel Attention (DCA) mechanism, which efficiently captures the interaction information of local channels in different directions, thereby enhancing the accuracy of the detection model for prohibited items. Additionally, a multi-scale fusion bypass (MFB) branch is designed after the backbone network to integrate information from feature maps at different layers, thereby mitigating the interference of scale variations on the model. Furthermore, the loss function is redesigned to enable the model to automatically adjust its focus on hard samples, improving the overall detection performance. Experimental results on the SIXray dataset demonstrate that the proposed model achieves a mean Average Precision (mAP) of 91.43%, representing a 9.17% improvement over the original algorithm, thereby validating the effectiveness of the proposed method.
Authors and Affiliations
ZHAOXiaotao, LIXinwei,
Comparative Study of Device Lifecycles in Different Healthcare Settings
This research targets the performances and lifetime of medical devices with regard to preventive maintenance (PM) practices in large and small health settings. A new maintenance effectiveness metric (MEM) has been develo...
Design and Development Performance Assessment System for Medical Staff Implementing the Regulations of the Indonesian Ministry of Health (Case Study: Arifin Achmad General Hospital, Riau Province)
The difficulty in conducting performance assessments of medical staff is the background for the Arifin Achmad Regional General Hospital (RSUD) to implement performance assessment standardization which refers to the techn...
An End to End Verifiability Proposal for the OTP-Vote Electronic Voting Model
Electronic voting systems have generated significant controversy in contemporary society due to the mistrust of the electorate. This mistrust is often rooted in the perceived inability to audit these systems with the sam...
Mass and Radius of a Fabriton Particle
In this research work, the mass and radius of a fabriton particle could be determined, and confirm that the subatomic particles and tremendous numbers of fabriton particles precisely fill the structure of atoms. Atom is...
Harmony in Nodes: Exploring Efficiency and Resilience in Distributed Systems
This research, titled "Harmony in Nodes: Exploring Efficiency and Resilience in Distributed Systems," delves into the intricate dynamics of distributed systems, aiming to uncover the delicate balance required for optimal...