Md Parvej Mosharaf

Mr. Md Parvej Mosharaf

orcid

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
4 September 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
SYsiyao yan
siyao yan
JZJing Zhang
Jing Zhang
MMMd Parvej Mosharaf
Md Parvej Mosharaf
TZTianchi Zhang
Tianchi Zhang
This paper proposes a novel Sea Drift Trajectory Prediction method based on the Quantum Convolutional Long Short-Term Memory (QCNN-LSTM) model. Accurately predicting sea drift trajectories is a challenging task, as they are influenced by various complex factors, such as ocean currents, wind speed, and wave morphology. Therefore, in a complex marine environment, there is a need for more applicable and computationally advanced prediction methods. Our approach combines quantized convolutional neural networks with Long Short-Term Memory networks, utilizing two different input types of prediction to enhance the network’s applicability. By incorporating quantization techniques, we improve the computational power and accuracy of the trajectory prediction. We evaluate our method using sea drift datasets and AUV drift trajectory datasets, comparing it with other commonly used traditional methods. The experimental results demonstrate significant improvements in accuracy and robustness achieved by our proposed Quantum Convolutional Long Short-Term Memory model. Regardless of the input mode employed, the accuracy consistently surpasses 98%. In conclusion, our research provides a new approach for sea drift trajectory prediction, enhancing prediction accuracy and providing valuable insights for marine environmental management and related decision-making. Future research can further explore and optimize this model to have a greater impact on marine prediction and applications.

10 May 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
LQLi Qin
Li Qin
TZTianchi Zhang
Tianchi Zhang
MMMd Parvej Mosharaf
Md Parvej Mosharaf
JZJing Zhang
Jing Zhang
Colorectal cancer is one of the three most common cancers worldwide. Polyps are complex and have unclear boundaries, often leading to inaccurate boundary segmentation and missed detections. To address these challenges, we propose a boundary constraint multi-level feature aggregation framework called BMFA-Net to precisely segment polyps. The framework comprises four key modules. First, the parallel partial decoder is introduced to aggregate high-level features within the network to generate a globally informative semantic map serving as the initial guidance region for reverse erasing the foreground. Second, we propose an efficient atrous convolution attention module to effectively aggregate local and global contextual information over multiple levels. Additionally, a multi-level feature aggregation mechanism is designed and placed among the efficient atrous convolution attention modules to enable the network to capture a large amount of semantic structure while preserving intricate details. Finally, a boundary constraint reverse attention module is proposed to perform the boundary constraint while removing the foreground to improve the quality of boundary segmentation. Extensive experiments demonstrated the superiority and versatility of our framework compared with state-of-the-art methods; specifically, it achieved a mean Dice score of 0.922 on the CVC-300 dataset.

11 May 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
PLPeng Li
Peng Li
SGShufang Guoshufang
Shufang Guoshufang
CZChenghao Zhang
Chenghao Zhang
MMMd Parvej Mosharaf
Md Parvej Mosharaf
JZJing Zhang
Jing Zhang
The rapid development of high-throughput technology has generated a large amount of protein–protein interaction (PPI) data, which provide a large amount of data support for constructing dynamic protein–protein interaction networks (PPINs). Constructing dynamic PPINs and applying them to recognize protein complexes has become a hot research topic. Most existing methods for complex recognition cannot fully mine the information of PPINs. To address this problem, we propose a construction method of dynamic weighted protein network by multi-level embedding (DWPNMLE). It can reflect the protein network’s dynamics and the protein network’s higher-order proximity. Firstly, the protein active period is calculated to divide the protein subnetworks at different time points. Then, the connection probability is used for the proteins possessing the same time points to judge whether there is an interaction relationship between them. Then, the corresponding protein subnetworks (multiple adjacency matrices) are constructed. Secondly, the multiple feature matrices are constructed using one-hot coding with the gene ontology (GO) information. Next, the first embedding is performed using variational graph auto-encoders (VGAEs) to aggregate features efficiently, followed by the second embedding using deep attributed network embedding (DANE) to strengthen the node representations learned in the first embedding and to maintain the first-order and higher-order proximity of the original network; finally, we compute the cosine similarity to obtain the final dynamic weighted PPIN. To evaluate the effectiveness of DWPNMLE, we apply four classical protein-complex-recognition algorithms on the DWPNMLE and compare them with two other dynamic protein network construction methods. The experimental results demonstrate that DWPNMLE significantly enhances the accuracy of complex recognition with high robustness, and the algorithms’ efficiency is also within a reasonable range.

13 September 2023 in COVID
Verified
Journal Article
Open access
PDF file
PLPeng Li
Peng Li
MMMd Parvej Mosharaf
Md Parvej Mosharaf
CZChenghao Zhang
Chenghao Zhang
SGShufang Guoshufang
Shufang Guoshufang
JZJing Zhang
Jing Zhang
In bioinformatics research, traditional machine-learning methods have demonstrated efficacy in addressing Euclidean data. However, real-world data often encompass non-Euclidean forms, such as graph data, which contain intricate structural patterns or high-order relationships that elude conventional machine-learning approaches. Representation learning seeks to derive valuable data representations from enhancing predictive or analytic tasks, capturing vital patterns and structures. This method has proven particularly beneficial in bioinformatics and biomedicine, as it effectively handles high-dimensional and sparse data, detects complex biological patterns, and optimizes predictive performance. In recent years, graph representation learning has become a popular research topic. It involves the embedding of graphs into a low-dimensional space while preserving the structural and attribute information of the graph, enabling better feature extraction for downstream tasks. This study extensively reviews representation learning advancements, particularly in the research of representation methods since the emergence of COVID-19. We begin with an analysis and classification of neural-network-based language model representation learning techniques as well as graph representation learning methods. Subsequently, we explore their methodological innovations in the context of COVID-19, with a focus on the domains of drugs, public health, and healthcare. Furthermore, we discuss the challenges and opportunities associated with graph representation learning. This comprehensive review presents invaluable insights for researchers as it documents the development of COVID-19 and offers experiential lessons to preempt future infectious diseases. Moreover, this study provides guidance regarding future bioinformatics and biomedicine research methodologies.

17 April 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MMMd Parvej Mosharaf
Md Parvej Mosharaf
JSJie Su
Jie Su
JZJing Zhang
Jing Zhang
Accurate brain tumor segmentation in magnetic resonance imaging (MRI) is essential for diagnosis, treatment planning, and therapy monitoring. Conventional deep learning models often struggle with large variations in tumor shape, size, and contrast, as well as severe foreground–background imbalance. To address these challenges, this study presents Deep-MiSR, an enhanced encoder–decoder framework built upon DeepLabV3+ with a MobileNetV2 backbone, tailored for single-modality contrast-enhanced T1-weighted (T1CE) MRI segmentation. Three complementary components are integrated into the architecture: mixed depthwise convolution (MixConv) with heterogeneous kernels within the atrous spatial pyramid pooling module for multi-scale feature aggregation, a squeeze-and-excitation block for adaptive channel recalibration, and R-Drop regularization that enforces prediction consistency via symmetric Kullback–Leibler divergence. The model was evaluated on 3064 T1CE slices from 233 patients drawn from the publicly available Nanfang Hospital brain MRI dataset. Deep-MiSR achieved a Dice similarity coefficient of 0.9281, a mean intersection-over-union of 0.8738, a precision of 0.8839, and a 95th-percentile Hausdorff distance of 7.69 mm, demonstrating consistent improvements over both the DeepLabV3+ baseline and all prior methods evaluated on the same data. Ablation studies confirmed that each component contributes independently, with R-Drop providing the largest individual gain. These findings demonstrate that combining multi-scale convolution, channel attention, and consistency regularization constitutes an effective and computationally practical strategy for robust single-modality brain tumor segmentation.