סמינר בבית הספר להנדסת חשמל ומחשבים
סמינר מחלקתי
10:00
Nadav Simor. (M.Sc. Student ) Supervisor : Prof. Yossi Rosen
Quantitative wave front phase imaging by holography with a phase-shifting method
Abstract:
Quantitative phase imaging enables high-resolution analysis of transparent samples without the need for staining, making it valuable across diverse fields, such as biology and astronomy. In this work, we introduce a phase-shifting self-reference holography method designed for single-plane quantitative phase imaging. This approach aims to improve phase accuracy while maintaining system simplicity. By combining optical manipulation with a phase-shifting algorithm, we reconstruct high-fidelity phase maps from a single hologram. The simulation results validate the method’s ability to reconstruct precise phase information using three or four temporal shots. These results can support the development of more compact and real-time quantitative phase imaging systems for various industrial applications
10:25
Ari Granevich (M.Sc. Student)
Title - Dual-head diffusion model for range-Doppler target detection with heavy tailed clutter
Advisors: Joseph Tabrikian, Haim Permuter, Igal Bilik.
This work introduces a dual-head diffusion model for multi-target detection in radar range-Doppler maps containing heavy-tailed, correlated clutter. The architecture employs a diffusion-based generative framework, featuring a denoising backbone that reconstructs noisy range-Doppler maps and a parallel detection head that pinpoints target locations within the cleaned representations. We employ an end-to-end training scheme enabling shared feature learning between denoising and detection branches, enhancing robustness to non-Gaussian clutter. The performance of the proposed architecture was evaluated using simulated radar datasets with varied clutter correlation and target densities in terms of probability of detection (Pd) and probability of false alarm (Pfa), and compared to CA-CFAR, TM-CFAR, and neural network detectors such as DAFC. The dual-head diffusion model consistently achieves higher Pd compared to baselines, demonstrating superior detection accuracy and resilience under challenging clutter conditions. These results highlight the potential of diffusion-based methods to advance multi-target radar detection in complex environments
10:50
Amro Asali (MS. C Student)
Super Visors: Yehuda Ben-Shimol, Itzchak Lapidot
Abstract: The objective of automatic speaker verification (ASV) systems
is to determine whether a given test speech utterance corresponds to a claimed enrolled speaker. These systems have a wide range of applications, and ensuring their reliability is crucial. In this paper, we propose a spoofing-robust automatic speaker verification (SASV) system employing a score-aware
gated attention (SAGA) fusion scheme, integrating scores from a pre-trained countermeasure (CM) with speaker embeddings from a pre-trained ASV. Specifically, we employ the AASIST
and ECAPA-TDNN models. SAGA acts as an adaptive gating mechanism, where the CM score determines how strongly ASV
embeddings influence the final SASV decision.
11:15
Bar Hen (M.Sc. student)
Under the supervision of Prof. Gil Shalev
Abstract
The early detection of bacterial contamination in a rapid, specific, and with minimal sample processing is a critical unmet need for global food safety and infectious disease prevention. Escherichia coli (E. coli), a widely used indicator of faecal contamination, poses significant health risks in its pathogenic forms and continues to challenge conventional detection methods, which are often slow, labour-intensive, and incompatible with complex samples. The study reports the Meta-Nano-Channel biological Field-Effect Transistor (MNC bioFET) for real-time, specific, label-free, quantitative and real-time sensing of E. coli in 0.5 µL of undiluted Lysogenic Broth (LB). The MNC bioFET is a fabricated in a complementary metal-oxide-semiconductor (CMOS) silicon-on-insulator (SOI) technology and biofunctionalized with monoclonal anti-E. coli O111:B4 LPS antibodies. I demonstrate E. coli sensing with a limit-of-detection (LOD) of single bacteria and with excellent sensitivity and linearity. Extensive control experiments and physical surface characterizations are performed. This work presents a breakthrough in E. coli biosensing and provides a solid foundation for future research on real-time microbial diagnostics.
Keywords: E. coli sensing, BioFET, Real-time bacterial detection, Single-cell analysis, Label-free biosensor, Biological activity monitoring
