Meng Xu

Meng Xu

Auto-generated profile, based on publicly available information

Meng Xu

Kean University

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 January 2023 in IEEE Access
Verified
Journal Article
Open access
MXMeng Xu
Meng Xu
Kuan HuangKuan Huang
Kuan Huang
XQXiaojun Qi
Xiaojun Qi
Breast ultrasound (BUS) imaging is commonly used in the early detection of breast cancer as a portable, valuable, and widely available diagnosis tool. Automated BUS image classification and segmentation can assist radiologists in making accurate and fast decisions. Recent studies illustrate that tumor, peritumoral, and background regions of BUS images provide valuable information for BUS image segmentation or classification. However, few studies have investigated the influence of these three regions on multi-task learning. In this study, we propose an RMTL-Net to simultaneously segment tumor regions and classify tumors in BUS images into benign or malignant categories. To improve both segmentation and classification performance, we design a regional attention (RA) module that employs the predicted probability maps to automatically guide the classifier to learn important category-sensitive information in the tumor, peritumoral, and background regions and seamlessly fuse them to obtain a better feature representation. We conduct detailed ablation experiments of the proposed RA module and comparative experiments with four recent state-of-the-art peer multi-task learning methods, three single-task segmentation methods, and four single-task classification methods on two public BUS datasets. Experimental results show that the proposed RMTL-Net achieves the best overall segmentation and classification accuracy in terms of five segmentation metrics and six classification metrics.

12 March 2024 in IEEE Access
Verified
Journal Article
Open access
MXMeng Xu
Meng Xu
JHJianhua Huang
Jianhua Huang
Kuan HuangKuan Huang
Kuan Huang
FLFeifei Liu
Feifei Liu
Breast Ultrasound (BUS) imaging is an essential tool for the early detection of breast cancer. The Breast Imaging Reporting and Data System (BI-RADS) in BUS images helps standardize the interpretation and reporting process by categorizing breast tumors into multiple classes, which enables radiologists to make more accurate diagnoses and treatment plans. However, most existing classification methods distinguish only between benign and malignant categories. In addition, features extracted by classic convolutional neural networks tend to be insufficient when subdividing BUS images into fine-grained BI-RADS classes, as they typically do not consider prior knowledge in medical applications, such as foreground shape. To address the above problems, we propose a novel fine-grained BI-RADS classification approach that integrates tumor edges to provide more efficient discriminative features. Firstly, weakly supervised pseudo-label generation: we detect coarse tumor edge regions utilizing a pre-trained PiDiNet and two novel loss functions based on prior knowledge from our dataset. The detected tumor edges are subsequently used as pseudo-labels for the next step. Secondly, co-training a tumor edge detection network and a BI-RADS classification network: edge images generated by the edge detection network are used as weight masks to highlight tumor edge regions as discriminative parts for better classification results, especially for categories with high similarities. The proposed method is evaluated on a BUS image dataset of 1061 images with BI-RADS categories. Experimental results indicate that the proposed method significantly improves over the baseline model by 4.73% in terms of top-1 accuracy.


4 February 2025 in IEEE Access
Verified
Journal Article
Open access
BHBIN HU
BIN HU
JYJunyong You
Junyong You
Kuan HuangKuan Huang
Kuan Huang
MXMeng Xu
Meng Xu
DLDan Liu
Dan Liu
SMSugang Ma
Sugang Ma
Deep neural networks (DNNs) have proven highly effective in image classification and object detection, with critical applications such as autonomous driving, security, Unmanned Aerial Vehicle (UAV) obstacle avoidance, and robot vision. However, the sensitivity of both input data and DNN architectures has increased the demand for secure execution on mobile devices. To address these concerns, hardware-based Trusted Execution Environments (TEEs), such as ARM TrustZone, are employed to ensure secure DNN operations. Despite their potential, deploying full-scale DNNs within mobile TEEs is challenging due to stringent resource and performance constraints. This work presents a novel security framework to enable the secure and efficient execution of DNNs within mobile TEEs. Specifically, the framework employs an effective dynamic suppression method that identifies and prunes unimportant neurons, resulting in a more compact and computationally efficient network model. Foundational neuron restructuring optimizes memory usage by deallocating storage for pruned neurons. Additionally, a dynamic slicing method combined with composite layer integration and loading strategies partitions pruned DNN models into multiple segments designed to fit within TEE memory constraints. Experiments conducted across various DNN models demonstrate statistically significant performance improvements. On the ImageNet dataset, EffTEE achieved an average Weight Compression Ratio (CR) of 82:3% ± 1:2%, a FLOPs CR of 85:7% ± 1:0%, and a 33:2% ± 1:4% reduction in inference time while maintaining97:2% accuracy. On CIFAR-10, EffTEE reduced inference time by 29:6%±1:2% with accuracy retention within ±0:8% of the original model. For object detection on the MS COCO dataset, EffTEE reduced inference time by 35:1%±1:5%, achieving consistent improvements in Frames Per Second (FPS). Overall, EffTEE achieves a significant 2 to 60 times reduction in inference time while maintaining accuracy comparable to existing methods for securing entire DNNs within mobile TEEs. Statistical validation, with a mean accuracy retention of 97:2% ± 0:8% and a p-value of <0.05, confirms the framework's robustness. These results highlight EffTEE's capability for robust and efficient DNN deployment in resource-constrained environments, balancing security, performance, and accuracy.


10 June 2025 in IEEE Transactions on Artificial Intelligence
Verified
Journal Article
Closed access
Kuan HuangKuan Huang
Kuan Huang
MXMeng Xu
Meng Xu
YWYingfeng Wang
Yingfeng Wang
The success of adversarial attack methods suggests a small input change may mislead a trained machine-learning model. For example, changing one pixel of an image may cause the trained model to misclassify this updated image. Uncertainty quantification is crucial for detecting misclassifications; hence, precise uncertainty quantification, meaning uncertainty estimates that closely align with prediction correctness, is essential. We assume that misclassified samples should exhibit high uncertainty while correctly classified samples should exhibit low uncertainty. To evaluate the performance of uncertainty quantification, we investigate the task of uncertainty-based misclassification detection under adversarial attack conditions. Our findings suggest that existing uncertainty quantification methods are unable to accurately identify misclassified predictions resulting from adversarial attacks due to training issues. We propose a simple adversarial training strategy for improving uncertainty quantification. Our results show that adversarial training improves the reliability of uncertainty quantification by better aligning uncertainty with prediction correctness. Specifically, we observe consistent improvements in misclassification detection performance, measured by AUROC and AUPR, across clean and adversarial samples. The code is available at https://github.com/kuanhuang0624/AdvUQ.