Mr. Shajib Ghosh Profile

Shajib Ghosh

at Univ of Florida

SPIE Involvement:

Author

Publications (8)

Proceedings Article | 17 October 2024 Presentation + Paper

Multi-modal printed circuit board netlist extraction with x-ray and optical imaging

Patrick Craig, Nitin Varshney, Antika Roy, Shajib Ghosh, Chandraman Patil, Hamed Dalir, Navid Asadizanjani

Proceedings Volume 13152, 131520Q (2024) https://doi.org/10.1117/12.3027170

KEYWORDS: X-rays, X-ray imaging, Data modeling, X-ray optics, Reverse modeling, Optical imaging, Machine learning, Image segmentation, Design, Printing

Read Abstract +

Proceedings Article | 4 October 2024 Poster

Exploring physics-informed machine learning for system matrix formulation in x-ray imaging forward models

Shajib Ghosh, Nitin Varshney, Md Mahfuz Al Hasan, Antika Roy, Patrick Craig, Sanjeev Koppal, Hamed Dalir, Navid Asadizanjani

Proceedings Volume 13152, 131521W (2024) https://doi.org/10.1117/12.3027381

KEYWORDS: Machine learning, X-ray imaging, Matrices, Imaging systems, Medical imaging, Mathematical modeling, Inspection, Electronic components, Data modeling, Algorithm development

Read Abstract +

This study presents an innovative approach to constructing a representative system matrix in x-ray imaging forward models. The approach leverages the combination of machine learning algorithms and fundamental physical principles through the use of physics-informed machine learning (PIML). The main goal is to seamlessly integrate machine learning algorithms with core physical principles to provide a nuanced perspective on the development of an interpretable and adaptive system matrix. In contrast to traditional data-intensive methods, this research intentionally prioritizes the incorporation of physics-based constraints into the machine learning framework. The methodology involves carefully extracting relevant features from x-ray imaging data to capture essential object characteristics, which are then integrated into a machine learning model. By including physics-based constraints, the model aligns with the underlying principles that govern x-ray interactions. Through rigorous mathematical validation and preliminary experimentation, the approach demonstrates its feasibility, particularly in situations where acquiring extensive datasets is challenging. From a technical standpoint, the strength of this methodology lies in the inherent adaptability and interpretability of the system matrix, which are crucial for accurate image reconstruction and measurement prediction. The implications of this research span diverse domains and highlight the potential transformative effects on x-ray imaging applications in electronics, medical imaging, and material inspection. In the realm of electronics, the adaptable system matrix improves non-destructive testing by aiding in defect detection and ensuring the reliability of electronic components. In medical imaging, enhanced interpretability leads to improved diagnostic accuracy while reducing radiation exposure. In material inspection, this approach facilitates the identification of structural anomalies and material composition, thereby advancing quality control practices. While recognizing the preliminary nature of the framework, this study lays the groundwork for future research at the intersection of machine learning and physics in x-ray imaging, representing a progressive step towards unlocking transformative possibilities for enhanced accuracy and adaptability across various domains.

Proceedings Article | 4 October 2024 Poster

Challenges and opportunities in non-destructive characterization of stacked IC packaging: insights from SAM and 3D x-ray analysis

Nitin Varshney, Shajib Ghosh, Patrick Craig, Himanandhan Reddy Kottur, Hamed Dalir, Navid Asadizanjani

Proceedings Volume 13152, 1315225 (2024) https://doi.org/10.1117/12.3027317

KEYWORDS: Nondestructive evaluation, X-rays, X-ray characterization, Packaging, Advanced packaging, 3D image processing, X-ray imaging, Semiconductors, Reliability, Inspection

Read Abstract +

Proceedings Article | 3 October 2024 Presentation + Paper

Fault-marking: defect-pattern leveraged inherent fingerprinting of advanced IC package with thermoreflectance imaging

Nitin Varshney, Mohammad Shafkat Khan, Shajib Ghosh, Antika Roy, Je-Hyeong Bahk, Navid Asadizanjani

Proceedings Volume 13145, 131450A (2024) https://doi.org/10.1117/12.3027447

KEYWORDS: Thermoreflectance, Manufacturing, Digital watermarking, Reliability, Integrated circuits, Information security, Temperature metrology, Semiconductors, Reflection, Inspection

Read Abstract +

In the quest to secure the authenticity and ownership of advanced integrated circuit (IC) packages, a novel approach has been introduced in this paper that capitalizes on the inherent physical discrepancies within these components. This method, distinct from traditional strategies like physical unclonable functions (PUFs) and cryptographic techniques, harnesses the unique defect patterns naturally occurring during the manufacturing process. By employing thermo-reflectance imaging (TRI), a non-destructive evaluation technique, in this proposed method we inspect, characterize and localize defects within IC package structures such as Through-silicon Vias (TSV) and micro-bumps. TRI’s ability to detect minute temperature variations caused by defects enables the creation of a detailed map that outlines the specific locations and types of manufacturing irregularities. This novel technique leverages the uniqueness of each IC’s defect pattern to generate an inherent identifier or ’fault-mark.’ These identifiers are derived from the specific arrangement and combination of defects, making them virtually impossible to replicate or forge due to the randomness and complexity of the manufacturing process variations. The creation of these fault-marks offers a robust and tamper-resistant means of authentication, providing a reliable method for establishing proof of ownership for advanced IC packages. The implementation of this approach not only can enhance supply chain security but also acts as a deterrent against the counterfeiting of IC packages. By verifying the authenticity of ICs against a reference database of fingerprints captured during the post-silicon validation stage, stakeholders can ensure the integrity of their components. This method’s potential of using inherent fingerprinting for reliable authentication and traceability of advanced IC packages is also been discussed, thereby offering a promising solution to the challenges of counterfeiting and unauthorized reproduction in the electronics industry.

Proceedings Article | 3 October 2024 Presentation + Paper

SHIELD-PCB: securing hardware inspection with enhanced learning and defense against adversarial examples in printed circuit boards images

Shajib Ghosh, Antika Roy, Nitin Varshney, Patrick Craig, Md Mahfuz Al Hasan, Sanjeev Koppal, Hamed Dalir, Navid Asadi Zanjani

Proceedings Volume 13138, 131380O (2024) https://doi.org/10.1117/12.3027383

KEYWORDS: Inspection, Adversarial training, Education and training, Data modeling, Visual process modeling, Performance modeling, Computer vision technology, Defense and security, Systems modeling, Process modeling

Read Abstract +

This study aims to investigate the potential of enhancing the resilience of computer vision systems in the context of intelligent Printed Circuit Board (PCB) inspection through the integration of techniques that filter out adversarial examples. PCBs, which are crucial components of electronic devices, require reliable inspection methods. However, current computer vision models are vulnerable to adversarial attacks that can compromise their accuracy. Our research introduces an evolving approach that combines advanced deep learning architectures with adversarial training methods. The initial steps involve training a robust PCB inspection model using a diverse dataset and generating adversarial examples through carefully designed perturbations. Subsequently, the model is exposed to these adversarial examples during a dedicated training phase, enabling it to adapt to variations introduced by potential adversaries. To counter the impact of adversarial examples on classification decisions during real-time inspections, a filtration mechanism is implemented to identify and discard them. Preliminary experimentation and ongoing evaluations demonstrate promising progress in enhancing the resilience of PCB inspection models against adversarial attacks. Although the filtration mechanism is still in its early stages, it shows potential in identifying and neutralizing potential threats, contributing to efforts aimed at strengthening the reliability and trustworthiness of inspection outcomes. Moreover, the adaptability of the proposed methodology to various PCB designs, including different components, orientations, and lighting conditions, indicates the potential for transformative advancements in computer vision systems in critical domains. This research underscores the need for continued investigation into the evolving landscape of adversarial example filtration, presenting a potential avenue for fortifying intelligent inspection systems against adversarial threats in PCB inspection and beyond.

Proceedings Article | 3 October 2024 Presentation + Paper

Optimizing deep learning performance in PCB x-ray inspection through synthetic data tuner

Shajib Ghosh, Md Mahfuz Al Hasan, Patrick Craig, Antika Roy, Nitin Varshney, Sanjeev Koppal, Hamed Dalir, Navid Asadizanjani

Proceedings Volume 13138, 131380N (2024) https://doi.org/10.1117/12.3027380

KEYWORDS: Data modeling, Performance modeling, Education and training, RGB color model, X-rays, Gallium nitride, Inspection, Deep learning, Image classification, X-ray imaging

Read Abstract +

In the domain of printed circuit board (PCB) X-ray inspection, the effectiveness of deep learning models greatly depends on the availability and quality of annotated data. The utilization of data augmentation techniques, particularly through the utilization of synthetic data, has emerged as a promising strategy to improve model performance and alleviate the burden of manual annotation. However, a significant question remains unanswered: What is the optimal amount of synthetic data required to effectively augment the dataset and enhance model performance? This study introduces the Synthetic Data Tuner, a comprehensive framework developed to address this crucial question and optimize the performance of deep learning models for PCB X-ray inspection tasks. By employing a combination of cutting-edge deep learning architectures and advanced data augmentation techniques, such as generative adversarial networks (GANs) and variational autoencoders (VAEs), the Synthetic Data Tuner systematically assesses the impact of different levels of synthetic data integration on model accuracy, robustness, and generalization. Through extensive experimentation and rigorous evaluation procedures, our results illustrate the intricate relationship between the quantity of synthetic data and model performance. We elucidate the phenomenon of diminishing returns, where model performance reaches a saturation point beyond a specific threshold of synthetic data augmentation. Moreover, we determine the optimal balance between synthetic and real data, achieving a harmonious equilibrium that maximizes performance improvements while mitigating the risk of overfitting. Additionally, our findings emphasize the significance of data diversity and quality in the generation of synthetic data, highlighting the importance of domain-specific knowledge and context-aware augmentation techniques. By providing insights into the complex interplay between synthetic data augmentation and deep learning model performance, the Synthetic Data Tuner not only advances the current state-of-the-art in PCB X-ray inspection but also offers valuable insights and methodologies applicable to various computer vision and industrial inspection domains.

Proceedings Article | 3 October 2024 Presentation + Paper

Optimizing the segment anything model for PCB component segmentation in x-ray images through few-shot parameter-efficient fine-tuning

Antika Roy, Md Mahfuz Al Hasan, Shajib Ghosh, Nitin Varshney, Patrick Craig, Charles Woychik, Reza Forghani, Navid Asadizanjani

Proceedings Volume 13138, 131380P (2024) https://doi.org/10.1117/12.3027646

KEYWORDS: Image segmentation, Data modeling, Education and training, Performance modeling, X-rays, X-ray imaging, Design, Reverse modeling, Machine learning, Visual process modeling

Read Abstract +

Proceedings Article | 1 October 2024 Presentation + Paper

Terahertz time-domain spectroscopy (THz-TDS) fingerprinting for integrated circuit (IC) identification in tracking and tracing applications

Patrick Craig, Chengjie Xi, Nitin Varshney, Wiliam Mitchell, Shajib Ghosh, Navid Asadizanjani

Proceedings Volume 13141, 131410B (2024) https://doi.org/10.1117/12.3027511

KEYWORDS: Terahertz radiation, Terahertz spectroscopy, Spectroscopy, Statistical modeling, Integrated circuits, Near field, Statistical analysis, Random forests, Pulse signals, Packaging

Read Abstract +

Showing 5 of 8 publications

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years