5.
Mattia Litrico; Alessio Del Bue; Pietro Morerio
Guiding Pseudo-labels with Uncertainty Estimation for Source-free Unsupervised Domain Adaptation Inproceedings
In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023.
@inproceedings{litrico_2023_CVPR,
title = {Guiding Pseudo-labels with Uncertainty Estimation for Source-free Unsupervised Domain Adaptation},
author = {Mattia Litrico and Alessio Del Bue and Pietro Morerio},
url = {https://arxiv.org/abs/2303.03770
https://github.com/MattiaLitrico/Guiding-Pseudo-labels-with-Uncertainty-Estimation-for-Source-free-Unsupervised-Domain-Adaptation},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
abstract = {Standard Unsupervised Domain Adaptation (UDA) methods assume the availability of both source and target data during the adaptation. In this work, we investigate Source-free Unsupervised Domain Adaptation (SF-UDA), a specific case of UDA where a model is adapted to a target domain without access to source data. We propose a novel approach for the SF-UDA setting based on a loss reweighting strategy that brings robustness against the noise that inevitably affects the pseudo-labels. The classification loss is reweighted based on the reliability of the pseudo-labels that is measured by estimating their uncertainty. Guided by such reweighting strategy, the pseudo-labels are progressively refined by aggregating knowledge from neighbouring samples. Furthermore, a self-supervised contrastive framework is leveraged as a target space regulariser to enhance such knowledge aggregation. A novel negative pairs exclusion strategy is proposed to identify and exclude negative pairs made of samples sharing the same class, even in presence of some noise in the pseudo-labels. Our method outperforms previous methods on three major benchmarks by a large margin. We set the new SF-UDA state-of-the-art on VisDA-C and DomainNet with a performance gain of +1.8% on both benchmarks and on PACS with +12.3% in the single-source setting and +6.6% in multi-target adaptation. Additional analyses demonstrate that the proposed approach is robust to the noise, which results in significantly more accurate pseudo-labels compared to state-of-the-art approaches.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Standard Unsupervised Domain Adaptation (UDA) methods assume the availability of both source and target data during the adaptation. In this work, we investigate Source-free Unsupervised Domain Adaptation (SF-UDA), a specific case of UDA where a model is adapted to a target domain without access to source data. We propose a novel approach for the SF-UDA setting based on a loss reweighting strategy that brings robustness against the noise that inevitably affects the pseudo-labels. The classification loss is reweighted based on the reliability of the pseudo-labels that is measured by estimating their uncertainty. Guided by such reweighting strategy, the pseudo-labels are progressively refined by aggregating knowledge from neighbouring samples. Furthermore, a self-supervised contrastive framework is leveraged as a target space regulariser to enhance such knowledge aggregation. A novel negative pairs exclusion strategy is proposed to identify and exclude negative pairs made of samples sharing the same class, even in presence of some noise in the pseudo-labels. Our method outperforms previous methods on three major benchmarks by a large margin. We set the new SF-UDA state-of-the-art on VisDA-C and DomainNet with a performance gain of +1.8% on both benchmarks and on PACS with +12.3% in the single-source setting and +6.6% in multi-target adaptation. Additional analyses demonstrate that the proposed approach is robust to the noise, which results in significantly more accurate pseudo-labels compared to state-of-the-art approaches.
4.
Domenico Bonanni; Mattia Litrico; Waqar Ahmed; Pietro Morerio; Tiziano Cazzorla; Elisa Spaccapaniccia; Franca Cattani; Marcello Allegretti; Andrea Rosario Beccari; Alessio Del Bue; Franck Martin
A Deep Learning Approach to Optimize Recombinant Protein Production in Escherichia coli Fermentations Journal Article
In: Fermentation, vol. 9, no. 6, 2023, ISSN: 2311-5637.
@article{fermentation9060503,
title = {A Deep Learning Approach to Optimize Recombinant Protein Production in Escherichia coli Fermentations},
author = {Domenico Bonanni and Mattia Litrico and Waqar Ahmed and Pietro Morerio and Tiziano Cazzorla and Elisa Spaccapaniccia and Franca Cattani and Marcello Allegretti and Andrea Rosario Beccari and Alessio Del Bue and Franck Martin},
url = {https://www.mdpi.com/2311-5637/9/6/503},
doi = {10.3390/fermentation9060503},
issn = {2311-5637},
year = {2023},
date = {2023-01-01},
journal = {Fermentation},
volume = {9},
number = {6},
abstract = {Fermentation is a widely used process in the biotechnology industry, in which sugar-based substrates are transformed into a new product through chemical reactions carried out by microorganisms. Fermentation yields depend heavily on critical process parameter (CPP) values which need to be finely tuned throughout the process; this is usually performed by a biotech production expert relying on empirical rules and personal experience. Although developing a mathematical model to analytically describe how yields depend on CPP values is too challenging because the process involves living organisms, we demonstrate the benefits that can be reaped by using a black-box machine learning (ML) approach based on recurrent neural networks (RNN) and long short-term memory (LSTM) neural networks to predict real time OD600nm values from fermentation CPP time series. We tested both networks on an E. coli fermentation process (upstream) optimized to obtain inclusion bodies whose purification (downstream) in a later stage will yield a targeted neurotrophin recombinant protein. We achieved root mean squared error (RMSE) and relative error on final yield (REFY) performances which demonstrate that RNN and LSTM are indeed promising approaches for real-time, in-line process yield estimation, paving the way for machine learning-based fermentation process control algorithms.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Fermentation is a widely used process in the biotechnology industry, in which sugar-based substrates are transformed into a new product through chemical reactions carried out by microorganisms. Fermentation yields depend heavily on critical process parameter (CPP) values which need to be finely tuned throughout the process; this is usually performed by a biotech production expert relying on empirical rules and personal experience. Although developing a mathematical model to analytically describe how yields depend on CPP values is too challenging because the process involves living organisms, we demonstrate the benefits that can be reaped by using a black-box machine learning (ML) approach based on recurrent neural networks (RNN) and long short-term memory (LSTM) neural networks to predict real time OD600nm values from fermentation CPP time series. We tested both networks on an E. coli fermentation process (upstream) optimized to obtain inclusion bodies whose purification (downstream) in a later stage will yield a targeted neurotrophin recombinant protein. We achieved root mean squared error (RMSE) and relative error on final yield (REFY) performances which demonstrate that RNN and LSTM are indeed promising approaches for real-time, in-line process yield estimation, paving the way for machine learning-based fermentation process control algorithms.
3.
Dario Allegra; Mattia Litrico; Maria Ausilia Napoli Spatafora; Filippo Stanco; Giovanni Maria Farinella
Exploiting Egocentric Vision on Shopping Cart for Out-of-Stock Detection in Retail Environments Inproceedings
In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, pp. 1735-1740, 2021.
@inproceedings{Allegra_2021_ICCV,
title = {Exploiting Egocentric Vision on Shopping Cart for Out-of-Stock Detection in Retail Environments},
author = {Dario Allegra and Mattia Litrico and Maria Ausilia Napoli Spatafora and Filippo Stanco and Giovanni Maria Farinella},
url = {https://openaccess.thecvf.com/content/ICCV2021W/ACVR/html/Allegra_Exploiting_Egocentric_Vision_on_Shopping_Cart_for_Out-of-Stock_Detection_in_ICCVW_2021_paper.html},
year = {2021},
date = {2021-10-01},
urldate = {2021-10-01},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops},
pages = {1735-1740},
abstract = {Continuous detection and efficient monitoring of Out-Of-Stock (OOS) of products in retail environments is a key factor to improve stores profits. Traditional methods require labour-intensive human work dedicated to checking for products to refill raising the requirement of automatic solutions to detect OOS. In this work, we focus on the problem of OOS detection from an egocentric perspective proposing a new weak annotation of the EgoCart dataset. We benchmark the considered challenge employing a deep learning approach for the detection of OOS areas. Specifically, we train a Convolutional Neural Network (CNN) to predict attention maps useful to find OOS in retail areas and hence suggest the retail employers where to intervene. We evaluate results with both objective measures and a subjective analysis provided by human which has reviewed the obtained OOS attention maps. The achieved performance demonstrates that the proposed pipeline is promising to help the refilling process in the retail domain.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Continuous detection and efficient monitoring of Out-Of-Stock (OOS) of products in retail environments is a key factor to improve stores profits. Traditional methods require labour-intensive human work dedicated to checking for products to refill raising the requirement of automatic solutions to detect OOS. In this work, we focus on the problem of OOS detection from an egocentric perspective proposing a new weak annotation of the EgoCart dataset. We benchmark the considered challenge employing a deep learning approach for the detection of OOS areas. Specifically, we train a Convolutional Neural Network (CNN) to predict attention maps useful to find OOS in retail areas and hence suggest the retail employers where to intervene. We evaluate results with both objective measures and a subjective analysis provided by human which has reviewed the obtained OOS attention maps. The achieved performance demonstrates that the proposed pipeline is promising to help the refilling process in the retail domain.
2.
Mattia Litrico; Sebastiano Battiato; Sotirios A. Tsaftaris; Mario Valerio Giuffrida
Semi-Supervised Domain Adaptation for Holistic Counting under Label Gap Journal Article
In: Journal of Imaging, vol. 7, no. 10, 2021, ISSN: 2313-433X.
@article{jimaging7100198b,
title = {Semi-Supervised Domain Adaptation for Holistic Counting under Label Gap},
author = {Mattia Litrico and Sebastiano Battiato and Sotirios A. Tsaftaris and Mario Valerio Giuffrida},
url = {https://www.mdpi.com/2313-433X/7/10/198},
doi = {10.3390/jimaging7100198},
issn = {2313-433X},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {Journal of Imaging},
journal = {Journal of Imaging},
volume = {7},
number = {10},
abstract = {This paper proposes a novel approach for semi-supervised domain adaptation for holistic regression tasks, where a DNN predicts a continuous value y∈R given an input image x. The current literature generally lacks specific domain adaptation approaches for this task, as most of them mostly focus on classification. In the context of holistic regression, most of the real-world datasets not only exhibit a covariate (or domain) shift, but also a label gap—the target dataset may contain labels not included in the source dataset (and vice versa). We propose an approach tackling both covariate and label gap in a unified training framework. Specifically, a Generative Adversarial Network (GAN) is used to reduce covariate shift, and label gap is mitigated via label normalisation. To avoid overfitting, we propose a stopping criterion that simultaneously takes advantage of the Maximum Mean Discrepancy and the GAN Global Optimality condition. To restore the original label range—that was previously normalised—a handful of annotated images from the target domain are used. Our experimental results, run on 3 different datasets, demonstrate that our approach drastically outperforms the state-of-the-art across the board. Specifically, for the cell counting problem, the mean squared error (MSE) is reduced from 759 to 5.62; in the case of the pedestrian dataset, our approach lowered the MSE from 131 to 1.47. For the last experimental setup, we borrowed a task from plant biology, i.e., counting the number of leaves in a plant, and we ran two series of experiments, showing the MSE is reduced from 2.36 to 0.88 (intra-species), and from 1.48 to 0.6 (inter-species).},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper proposes a novel approach for semi-supervised domain adaptation for holistic regression tasks, where a DNN predicts a continuous value y∈R given an input image x. The current literature generally lacks specific domain adaptation approaches for this task, as most of them mostly focus on classification. In the context of holistic regression, most of the real-world datasets not only exhibit a covariate (or domain) shift, but also a label gap—the target dataset may contain labels not included in the source dataset (and vice versa). We propose an approach tackling both covariate and label gap in a unified training framework. Specifically, a Generative Adversarial Network (GAN) is used to reduce covariate shift, and label gap is mitigated via label normalisation. To avoid overfitting, we propose a stopping criterion that simultaneously takes advantage of the Maximum Mean Discrepancy and the GAN Global Optimality condition. To restore the original label range—that was previously normalised—a handful of annotated images from the target domain are used. Our experimental results, run on 3 different datasets, demonstrate that our approach drastically outperforms the state-of-the-art across the board. Specifically, for the cell counting problem, the mean squared error (MSE) is reduced from 759 to 5.62; in the case of the pedestrian dataset, our approach lowered the MSE from 131 to 1.47. For the last experimental setup, we borrowed a task from plant biology, i.e., counting the number of leaves in a plant, and we ran two series of experiments, showing the MSE is reduced from 2.36 to 0.88 (intra-species), and from 1.48 to 0.6 (inter-species).
1.
Oliver Giudice; Mattia Litrico; Sebastiano Battiato
Single Architecture and Multiple task deep Neural Network for Altered Fingerprint Analysis Inproceedings
In: 2020 IEEE International Conference on Image Processing (ICIP), pp. 813-817, 2020.
@inproceedings{9191094,
title = {Single Architecture and Multiple task deep Neural Network for Altered Fingerprint Analysis},
author = {Oliver Giudice and Mattia Litrico and Sebastiano Battiato},
doi = {10.1109/ICIP40778.2020.9191094},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {2020 IEEE International Conference on Image Processing (ICIP)},
pages = {813-817},
abstract = {Fingerprints are one of the most copious evidence in a crime scene and, for this reason, they are frequently used by law enforcement for identification of individuals. But fingerprints can be altered. “Altered fingerprints” refers to intentionally damage of the friction ridge pattern and they are often used by smart criminals in hope to evade law enforcement. We use a deep neural network approach training an Inception-v3 architecture. This paper proposes a method for detection of altered fingerprints, identification of types of alterations and recognition of gender, hand and fingers. We also produce activation maps that show which part of a fingerprint the neural network has focused on, in order to detect where alterations are positioned. The proposed approach achieves an accuracy of 98.21%, 98.46%, 92.52%, 97.53% and 92,18% for the classification of fakeness, alterations, gender, hand and fingers, respectively on the SO.CO.FING. dataset.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Fingerprints are one of the most copious evidence in a crime scene and, for this reason, they are frequently used by law enforcement for identification of individuals. But fingerprints can be altered. “Altered fingerprints” refers to intentionally damage of the friction ridge pattern and they are often used by smart criminals in hope to evade law enforcement. We use a deep neural network approach training an Inception-v3 architecture. This paper proposes a method for detection of altered fingerprints, identification of types of alterations and recognition of gender, hand and fingers. We also produce activation maps that show which part of a fingerprint the neural network has focused on, in order to detect where alterations are positioned. The proposed approach achieves an accuracy of 98.21%, 98.46%, 92.52%, 97.53% and 92,18% for the classification of fakeness, alterations, gender, hand and fingers, respectively on the SO.CO.FING. dataset.