Francesco Paissan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Francesco Paissan has a total of 279 co-authors. Here are the top co-authors by number of shared publications:
- Elisabetta Farella
Fondazione Bruno Kessler, Trento, Italy
5 shared publications - Davide Dalle Pezze
Department of Information Engineering, University of Padua, Padua, Italy
3 shared publications - Gian Antonio Susto
University of Padova, Padova, Italy
3 shared publications - Alberto Ancilotto
Fondazione Bruno Kessler, Trento, Italy
2 shared publications - Alessio Brutti
Fondazione Bruno Kessler Italy
2 shared publications - Diego Tagnani
INFN, Sezione di Roma Tre, Via della Vasca Navale 84, I-00146 Rome
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 16 January 2026
- Published by American Physical Society (APS) in Physical Review Letters
- Vol. 136 (2) , 022701
First Results on the Search for Lepton Number Violating Neutrinoless Double- β Decay with the LEGEND-200 Experiment
The LEGEND Collaboration is searching for neutrinoless double-beta ( 0 ν β β ) decay by operating high-purity germanium detectors enriched in Ge 76 in a low-background liquid argon environment. Building on key technological innovations from the GERmanium Detector Array (GERDA) experiment and the MAJORANA DEMONSTRATOR experiment, LEGEND-200 has performed a first 0 ν β β decay search based on 61.0 kg yr of data. Over half of this exposure comes from our highest performing detectors, including newly developed inverted-coaxial detectors, and is characterized by an estimated background level of 0.5 − 0.2 + 0.3 cts / ( keV ton yr ) in the 0 ν β β decay signal region. A combined analysis of data from GERDA, the MAJORANA DEMONSTRATOR, and LEGEND-200, characterized by a 90% confidence level exclusion sensitivity of 2.8 × 10 26 yr on the half-life of 0 ν β β decay, reveals no evidence for a signal and sets a new observed lower limit at T 1 / 2 0 ν > 1.9 × 10 26 yr (90% confidence level). Assuming the decay is mediated by Majorana neutrinos, this corresponds to an upper limit on the effective Majorana mass in the range m β β < 75 – 200 meV , depending on the adopted nuclear matrix element.
- 16 January 2026
- Published by American Physical Society (APS) in Physical Review Letters
- Vol. 136 (2) , 022701
Top Cited PapersOpen AccessFirst Results on the Search for Lepton Number Violating Neutrinoless Double- β Decay with the LEGEND-200 Experiment
The LEGEND Collaboration is searching for neutrinoless double-beta ( 0 ν β β ) decay by operating high-purity germanium detectors enriched in Ge 76 in a low-background liquid argon environment. Building on key technological innovations from the GERmanium Detector Array (GERDA) experiment and the MAJORANA DEMONSTRATOR experiment, LEGEND-200 has performed a first 0 ν β β decay search based on 61.0 kg yr of data. Over half of this exposure comes from our highest performing detectors, including newly developed inverted-coaxial detectors, and is characterized by an estimated background level of 0.5 − 0.2 + 0.3 cts / ( keV ton yr ) in the 0 ν β β decay signal region. A combined analysis of data from GERDA, the MAJORANA DEMONSTRATOR, and LEGEND-200, characterized by a 90% confidence level exclusion sensitivity of 2.8 × 10 26 yr on the half-life of 0 ν β β decay, reveals no evidence for a signal and sets a new observed lower limit at T 1 / 2 0 ν > 1.9 × 10 26 yr (90% confidence level). Assuming the decay is mediated by Majorana neutrinos, this corresponds to an upper limit on the effective Majorana mass in the range m β β < 75 – 200 meV , depending on the adopted nuclear matrix element.
- 12 May 2025
- Published by Springer Nature
- p. 279-294
Latent Distillation for Continual Object Detection at the Edge
- 12 May 2025
- Published by Springer Nature
- p. 279-294
Latent Distillation for Continual Object Detection at the Edge
- Open Access
PaSTe: Improving the Efficiency of Visual Anomaly Detection at the Edge
Visual Anomaly Detection (VAD) has gained significant research attention for its ability to identify anomalous images and pinpoint the specific areas responsible for the anomaly. A key advantage of VAD is its unsupervised nature, which eliminates the need for costly and time-consuming labeled data collection. However, despite its potential for real-world applications, the literature has given limited focus to resource-efficient VAD, particularly for deployment on edge devices. This work addresses this gap by leveraging lightweight neural networks to reduce memory and computation requirements, enabling VAD deployment on resource-constrained edge devices. We benchmark the major VAD algorithms within this framework and demonstrate the feasibility of edge-based VAD using the well-known MVTec dataset. Furthermore, we introduce a novel algorithm, Partially Shared Teacher-student (PaSTe), designed to address the high resource demands of the existing Student Teacher Feature Pyramid Matching (STFPM) approach. Our results show that PaSTe decreases the inference time by 25%, while reducing the training time by 33% and peak RAM usage during training by 76%. These improvements make the VAD process significantly more efficient, laying a solid foundation for real-world deployment on edge devices.
Open AccessPaSTe: Improving the Efficiency of Visual Anomaly Detection at the Edge
Visual Anomaly Detection (VAD) has gained significant research attention for its ability to identify anomalous images and pinpoint the specific areas responsible for the anomaly. A key advantage of VAD is its unsupervised nature, which eliminates the need for costly and time-consuming labeled data collection. However, despite its potential for real-world applications, the literature has given limited focus to resource-efficient VAD, particularly for deployment on edge devices. This work addresses this gap by leveraging lightweight neural networks to reduce memory and computation requirements, enabling VAD deployment on resource-constrained edge devices. We benchmark the major VAD algorithms within this framework and demonstrate the feasibility of edge-based VAD using the well-known MVTec dataset. Furthermore, we introduce a novel algorithm, Partially Shared Teacher-student (PaSTe), designed to address the high resource demands of the existing Student Teacher Feature Pyramid Matching (STFPM) approach. Our results show that PaSTe decreases the inference time by 25%, while reducing the training time by 33% and peak RAM usage during training by 76%. These improvements make the VAD process significantly more efficient, laying a solid foundation for real-world deployment on edge devices.
- Open Access
- 11 September 2024
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 182, 109097
SpeechBrain-MOABB: An open-source Python library for benchmarking deep neural networks applied to EEG signals
- 11 September 2024
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 182, 109097
Open AccessSpeechBrain-MOABB: An open-source Python library for benchmarking deep neural networks applied to EEG signals
- Open Access
Latent Distillation for Continual Object Detection at the Edge
While numerous methods achieving remarkable performance exist in the Object Detection literature, addressing data distribution shifts remains challenging. Continual Learning (CL) offers solutions to this issue, enabling models to adapt to new data while maintaining performance on previous data. This is particularly pertinent for edge devices, common in dynamic environments like automotive and robotics. In this work, we address the memory and computation constraints of edge devices in the Continual Learning for Object Detection (CLOD) scenario. Specifically, (i) we investigate the suitability of an open-source, lightweight, and fast detector, namely NanoDet, for CLOD on edge devices, improving upon larger architectures used in the literature. Moreover, (ii) we propose a novel CL method, called Latent Distillation~(LD), that reduces the number of operations and the memory required by state-of-the-art CL approaches without significantly compromising detection performance. Our approach is validated using the well-known VOC and COCO benchmarks, reducing the distillation parameter overhead by 74\% and the Floating Points Operations~(FLOPs) by 56\% per model update compared to other distillation methods.
Open AccessLatent Distillation for Continual Object Detection at the Edge
While numerous methods achieving remarkable performance exist in the Object Detection literature, addressing data distribution shifts remains challenging. Continual Learning (CL) offers solutions to this issue, enabling models to adapt to new data while maintaining performance on previous data. This is particularly pertinent for edge devices, common in dynamic environments like automotive and robotics. In this work, we address the memory and computation constraints of edge devices in the Continual Learning for Object Detection (CLOD) scenario. Specifically, (i) we investigate the suitability of an open-source, lightweight, and fast detector, namely NanoDet, for CLOD on edge devices, improving upon larger architectures used in the literature. Moreover, (ii) we propose a novel CL method, called Latent Distillation~(LD), that reduces the number of operations and the memory required by state-of-the-art CL approaches without significantly compromising detection performance. Our approach is validated using the well-known VOC and COCO benchmarks, reducing the distillation parameter overhead by 74\% and the Floating Points Operations~(FLOPs) by 56\% per model update compared to other distillation methods.
- Open Access
- 11 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (3) , 1-16
XimSwap: Many-to-Many Face Swapping for TinyML
The unprecedented development of deep learning approaches for video processing has caused growing privacy concerns. To ensure data analysis while maintaining privacy, it is essential to address how to protect individuals’ identities. One solution is to anonymize data at the source, avoiding the transmission or storage of information that could lead to identification. This study introduces XimSwap, a novel deep learning technique for real-time video anonymization, which can remove facial identification features directly on edge devices with minimal computational resources. Our approach offers a comprehensive solution that guarantees privacy by design. This novel method for implementing face-swapping ensures that the pose and expression of a target face remain unchanged and can be used on embedded devices with very limited computational resources. By incorporating style transfer layers into convolutional ones and optimizing the network’s operation, we achieved a reduction of over 98% in the required operations and parameters compared with state-of-the-art architectures. Our approach also significantly reduces RAM usage, making it possible to implement the anonymization process on tiny edge devices, including microcontrollers, such as the STM32H743.
- 11 May 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (3) , 1-16
Open AccessXimSwap: Many-to-Many Face Swapping for TinyML
The unprecedented development of deep learning approaches for video processing has caused growing privacy concerns. To ensure data analysis while maintaining privacy, it is essential to address how to protect individuals’ identities. One solution is to anonymize data at the source, avoiding the transmission or storage of information that could lead to identification. This study introduces XimSwap, a novel deep learning technique for real-time video anonymization, which can remove facial identification features directly on edge devices with minimal computational resources. Our approach offers a comprehensive solution that guarantees privacy by design. This novel method for implementing face-swapping ensures that the pose and expression of a target face remain unchanged and can be used on embedded devices with very limited computational resources. By incorporating style transfer layers into convolutional ones and optimizing the network’s operation, we achieved a reduction of over 98% in the required operations and parameters compared with state-of-the-art architectures. Our approach also significantly reduces RAM usage, making it possible to implement the anonymization process on tiny edge devices, including microcontrollers, such as the STM32H743.
- Open Access
- 1 September 2023
- Published by IOP Publishing in Journal of Instrumentation
- Vol. 18 (09) , P09007
The Front-End electronics for the liquid Argon instrumentation of the LEGEND-200 experiment
In this paper we provide a detailed technical description of the Front-End (FE) electronics for the liquid Argon instrumentation of the LEGEND-200 experiment, searching for the very rare, hypothetical neutrinoless double β decay process at the Italian Laboratori Nazionali del Gran Sasso. The design stems from the need to read out the silicon photo-multiplier response to the scintillation light in the liquid Argon with excellent single-photon resolution. The FE electronics is required to be placed far from the detectors to meet the experiment's radio-purity constraints. This constraint represents a challenge for a high signal-to-noise ratio. We address how this could be achieved in a stable way. The system was installed in July 2021 and has been commissioned with the rest of LEGEND-200, proving we could attain a very low overall level of electrical noise of 250 μV peak-to-peak.
- 1 September 2023
- Published by IOP Publishing in Journal of Instrumentation
- Vol. 18 (09) , P09007
Open AccessThe Front-End electronics for the liquid Argon instrumentation of the LEGEND-200 experiment
In this paper we provide a detailed technical description of the Front-End (FE) electronics for the liquid Argon instrumentation of the LEGEND-200 experiment, searching for the very rare, hypothetical neutrinoless double β decay process at the Italian Laboratori Nazionali del Gran Sasso. The design stems from the need to read out the silicon photo-multiplier response to the scintillation light in the liquid Argon with excellent single-photon resolution. The FE electronics is required to be placed far from the detectors to meet the experiment's radio-purity constraints. This constraint represents a challenge for a high signal-to-noise ratio. We address how this could be achieved in a stable way. The system was installed in July 2021 and has been commissioned with the rest of LEGEND-200, proving we could attain a very low overall level of electrical noise of 250 μV peak-to-peak.
- Open Access
Scaling strategies for on-device low-complexity source separation with Conv-Tasnet
Recently, several very effective neural approaches for single-channel speech separation have been presented in the literature. However, due to the size and complexity of these models, their use on low-resource devices, e.g. for hearing aids, and earphones, is still a challenge and established solutions are not available yet. Although approaches based on either pruning or compressing neural models have been proposed, the design of a model architecture suitable for a certain application domain often requires heuristic procedures not easily portable to different low-resource platforms. Given the modular nature of the well-known Conv-Tasnet speech separation architecture, in this paper we consider three parameters that directly control the overall size of the model, namely: the number of residual blocks, the number of repetitions of the separation blocks and the number of channels in the depth-wise convolutions, and experimentally evaluate how they affect the speech separation performance. In particular, experiments carried out on the Libri2Mix show that the number of dilated 1D-Conv blocks is the most critical parameter and that the usage of extra-dilation in the residual blocks allows reducing the performance drop.
Open AccessScaling strategies for on-device low-complexity source separation with Conv-Tasnet
Recently, several very effective neural approaches for single-channel speech separation have been presented in the literature. However, due to the size and complexity of these models, their use on low-resource devices, e.g. for hearing aids, and earphones, is still a challenge and established solutions are not available yet. Although approaches based on either pruning or compressing neural models have been proposed, the design of a model architecture suitable for a certain application domain often requires heuristic procedures not easily portable to different low-resource platforms. Given the modular nature of the well-known Conv-Tasnet speech separation architecture, in this paper we consider three parameters that directly control the overall size of the model, namely: the number of residual blocks, the number of repetitions of the separation blocks and the number of channels in the depth-wise convolutions, and experimentally evaluate how they affect the speech separation performance. In particular, experiments carried out on the Libri2Mix show that the number of dilated 1D-Conv blocks is the most critical parameter and that the usage of extra-dilation in the residual blocks allows reducing the performance drop.
- Open Access
Low-complexity acoustic scene classification in DCASE 2022 Challenge
This paper presents an analysis of the Low-Complexity Acoustic Scene Classification task in DCASE 2022 Challenge. The task was a continuation from the previous years, but the low-complexity requirements were changed to the following: the maximum number of allowed parameters, including the zero-valued ones, was 128 K, with parameters being represented using INT8 numerical format; and the maximum number of multiply-accumulate operations at inference time was 30 million. The provided baseline system is a convolutional neural network which employs post-training quantization of parameters, resulting in 46.5 K parameters, and 29.23 million multiply-and-accumulate operations (MMACs). Its performance on the evaluation data is 44.2% accuracy and 1.532 log-loss. In comparison, the top system in the challenge obtained an accuracy of 59.6% and a log loss of 1.091, having 121 K parameters and 28 MMACs. The task received 48 submissions from 19 different teams, most of which outperformed the baseline system.
Open AccessLow-complexity acoustic scene classification in DCASE 2022 Challenge
This paper presents an analysis of the Low-Complexity Acoustic Scene Classification task in DCASE 2022 Challenge. The task was a continuation from the previous years, but the low-complexity requirements were changed to the following: the maximum number of allowed parameters, including the zero-valued ones, was 128 K, with parameters being represented using INT8 numerical format; and the maximum number of multiply-accumulate operations at inference time was 30 million. The provided baseline system is a convolutional neural network which employs post-training quantization of parameters, resulting in 46.5 K parameters, and 29.23 million multiply-and-accumulate operations (MMACs). Its performance on the evaluation data is 44.2% accuracy and 1.532 log-loss. In comparison, the top system in the challenge obtained an accuracy of 59.6% and a log loss of 1.091, having 121 K parameters and 28 MMACs. The task received 48 submissions from 19 different teams, most of which outperformed the baseline system.