arXiv subjects

Mahdi Esmaeili

Publications and source records attributed to Mahdi Esmaeili.

5 recordsLinked to original sources

arXiv2026

An Innovative Framework for Breast Cancer Detection Using Pyramid Adaptive Atrous Convolution, Transformer Integration, and Multi-Scale Feature Fusion

Breast cancer is one of the most common cancers among women worldwide, and its accurate and timely diagnosis plays a critical role in improving treatment outcomes. This thesis presents an innovative framework for detecting malignant masses in mammographic images by integrating the Pyramid Adaptive Atrous Convolution (PAAC) and Transformer architectures. The proposed approach utilizes Multi-Scale Feature Fusion to enhance the extraction of features from benign and malignant tissues and combines Dice Loss and Focal Loss functions to improve the model's learning process, effectively reducing errors in binary breast cancer classification and achieving high accuracy and efficiency. In this study, a comprehensive dataset of breast cancer images from INbreast, MIAS, and DDSM was preprocessed through data augmentation and contrast enhancement and resized to 227x227 pixels for model training. Leveraging the Transformer's ability to manage long-range dependencies with Self-Attention mechanisms, the proposed model achieved high accuracy in detecting cancerous masses, outperforming foundational models such as BreastNet, DeepMammo, Multi-Scale CNN, Swin-Unet, and SegFormer. The final evaluation results for the proposed model include an accuracy of 98.5\%, sensitivity of 97.8\%, specificity of 96.3\%, F1-score of 98.2\%, and overall precision of 97.9\%. These metrics demonstrate a significant improvement over traditional methods and confirm the model's effectiveness in identifying cancerous masses in complex scenarios and large datasets. This model shows potential as a reliable and efficient tool for breast cancer diagnosis and can be effectively integrated into medical diagnostic systems.

cs.CV↗

arXiv2025

Technical Review on RF-Amplifiers for Quantum Computer Circuits: New Architectures of Josephson Parametric Amplifier

Josephson Parametric Amplifiers (JPAs) are key components in quantum information processing due to their ability to amplify weak quantum signals with near-quantum-limited noise performance. This is essential for applications such as qubit readout, quantum sensing, and communication, where signal fidelity and coherence preservation are critical. Unlike CMOS and HEMT amplifiers used in conventional RF systems, JPAs are specifically optimized for millikelvin (mK) cryogenic environments. CMOS amplifiers offer good integration but perform poorly at ultra-low temperatures due to high noise. HEMT amplifiers provide better noise performance but are power-intensive and less suited for mK operation. JPAs, by contrast, combine low power consumption with ultra-low noise and excellent cryogenic compatibility, making them ideal for quantum systems. The first part of this study compares these RF amplifier types and explains why JPAs are preferred in cryogenic quantum applications. The second part focuses on the design and analysis of JPAs based on both single Josephson junctions and junction arrays. While single-junction JPAs utilize nonlinear inductance for amplification, they suffer from gain compression, limited dynamic range, and sensitivity to fabrication variations. To overcome these challenges, this work explores JPA designs using Josephson junction arrays. Arrays distribute the nonlinear response, enhancing power handling, linearity, impedance tunability, and coherence while reducing phase noise. Several advanced JPA architectures are proposed, simulated, and compared using quantum theory and CAD tools to assess performance trade-offs and improvements over conventional designs.

quant-ph↗

arXiv2025

Design of Qubit Readout Circuit for Purcell Rate Suppressing and Nonclassicality Enhancing

The Purcell effect, a common issue in qubit-resonator systems leading to both fidelity and nonclassicality losses is studied while its suppression is achieved using a novel qubit readout circuit design. Our approach utilizes a unique coupling architecture in which, the qubit first interacts with a filter resonator before coupling to the readout resonator. This configuration enables precise control over the Purcell decay rate and ac Stark factor without impacting on measuring time. The mentioned factor is highly sensitive to the coupling strength between the readout resonator and the filter, meaning that the factor adjustment directly impacts the qubit state detection. A major advantage of this design is that tuning the resonator-filter coupling strength is relatively straightforward, offering flexibility in fine-tuning ac Stark factor. This work extensively analyzes the system using full quantum mechanical theory, deriving the total Hamiltonian and investigates mode dynamics via quantum Langevin equations. Key parameters influencing the nonclassicality of output signals are also explored through quantum correlation metrics, including symplectic eigenvalues, quantum discord, and classical discord. The main goal is to find any compromises exsting between the ac Stark factor increasing and the quantum correlation created in the readout circuit. By optimizing the critical factors, by which the ac Stark factor mainly affected, the proposed design not only improves the distinguishability of the qubit states but also ensures robust nonclassicality in the output signals. Results demonstrate the potential of the proposed system to bridge the gap between a high fidelity readout and quantum correlation preservation in scalable quantum architectures.

quant-ph↗

arXiv2023

Lung Cancer Detection from CT Scan Images based on Genetic-Independent Recurrent Deep Learning

Lung cancer is one of the prevalence diseases in the world which cause many deaths. Detecting early stages of lung cancer is so necessary. So, modeling and simulating some intelligent medical systems is an essential which can help specialist to accurately determine and diagnose the disease. So this paper contributes a new lung cancer detection model in CT images which use machine learning methods. There are three steps in this model: noise reduction (pre-processing), segmentation (middle-processing) and optimize segmentation for detect exact are of nodules. This article use some filters for noise reduction and then use Independent Recurrent Neural Networks (IndRNN) as deep learning methods for segmentation which optimize and tune by Genetic Algorithm. The results represented that the proposed method can detect exact area of nodules in CT images.

eess.IV↗

arXiv2010

Finding Sequential Patterns from Large Sequence Data

Data mining is the task of discovering interesting patterns from large amounts of data. There are many data mining tasks, such as classification, clustering, association rule mining, and sequential pattern mining. Sequential pattern mining finds sets of data items that occur together frequently in some sequences. Sequential pattern mining, which extracts frequent subsequences from a sequence database, has attracted a great deal of interest during the recent data mining research because it is the basis of many applications, such as: web user analysis, stock trend prediction, DNA sequence analysis, finding language or linguistic patterns from natural language texts, and using the history of symptoms to predict certain kind of disease. The diversity of the applications may not be possible to apply a single sequential pattern model to all these problems. Each application may require a unique model and solution. A number of research projects were established in recent years to develop meaningful sequential pattern models and efficient algorithms for mining these patterns. In this paper, we theoretically provided a brief overview three types of sequential patterns model.

cs.DB↗