Search

Time-resolved single photon imaging is a promising imaging modality characterized by the unique capability of timestamping the arrivals of single photons. Single-Photon Avalanche Diodes (SPADs) are the leading technology for implementing modern time-resolved pixels, suitable for passive imaging with asynchronous readout. However, they are currently limited to small sized arrays, thus there is a lack of datasets for passive time-resolved SPAD imaging, which in turn hinders research on this peculiar imaging data. In this paper we describe a realistic simulation process for SPAD imaging, which takes into account both the stochastic nature of photon arrivals and all the noise sources involved in the acquisition process of time-resolved SPAD arrays. We have implemented this simulator in a software prototype able to generate arbitrary-sized time-resolved SPAD arrays operating in passive mode. Starting from a reference image, our simulator generates a realistic stream of timestamped photon detections. We use our simulator to generate a time-resolved version of MNIST, which we make publicly available. Our dataset has the purpose of encouraging novel research directions in time-resolved SPAD imaging, as well as investigating the performance of CNN classifiers in extremely low-light conditions., Comment: 12 pages, 4 figures. Accepted for Workshop on Synthetic Data for Computer Vision at ECCV 2024

We present Kernel-QuantTree Exponentially Weighted Moving Average (KQT-EWMA), a non-parametric change-detection algorithm that combines the Kernel-QuantTree (KQT) histogram and the EWMA statistic to monitor multivariate data streams online. The resulting monitoring scheme is very flexible, since histograms can be used to model any stationary distribution, and practical, since the distribution of test statistics does not depend on the distribution of datastream in stationary conditions (non-parametric monitoring). KQT-EWMA enables controlling false alarms by operating at a pre-determined Average Run Length ($ARL_0$), which measures the expected number of stationary samples to be monitored before triggering a false alarm. The latter peculiarity is in contrast with most non-parametric change-detection tests, which rarely can control the $ARL_0$ a priori. Our experiments on synthetic and real-world datasets demonstrate that KQT-EWMA can control $ARL_0$ while achieving detection delays comparable to or lower than state-of-the-art methods designed to work in the same conditions., Comment: AALTD workshop at ECML 2024 (https://ecml-aaltd.github.io/aaltd2024/)

The seismocardiographic signal is a promising alternative to the traditional ECG in the analysis of the cardiac activity. In particular, the systolic complex is known to be the most informative part of the seismocardiogram, thus requiring further analysis. State-of-art solutions to detect the systolic complex are based on Deep Learning models, which have been proven effective in pioneering studies. However, these solutions have only been tested in a controlled scenario considering only clean signals acquired from users maintained still in supine position. On top of that, all these studies consider data coming from a single dataset, ignoring the benefits and challenges related to a cross-dataset scenario. In this work, a cross-dataset experimental analysis was performed considering also data from a real-world scenario. Our findings prove the effectiveness of a deep learning solution, while showing the importance of a personalization step to contrast the domain shift, namely a change in data distribution between training and testing data. Finally, we demonstrate the benefits of a multi-channels approach, leveraging the information extracted from both accelerometers and gyroscopes data.

Zero-shot anomaly detection (ZSAD) targets the identification of anomalies within images from arbitrary novel categories. This study introduces AdaCLIP for the ZSAD task, leveraging a pre-trained vision-language model (VLM), CLIP. AdaCLIP incorporates learnable prompts into CLIP and optimizes them through training on auxiliary annotated anomaly detection data. Two types of learnable prompts are proposed: static and dynamic. Static prompts are shared across all images, serving to preliminarily adapt CLIP for ZSAD. In contrast, dynamic prompts are generated for each test image, providing CLIP with dynamic adaptation capabilities. The combination of static and dynamic prompts is referred to as hybrid prompts, and yields enhanced ZSAD performance. Extensive experiments conducted across 14 real-world anomaly detection datasets from industrial and medical domains indicate that AdaCLIP outperforms other ZSAD methods and can generalize better to different categories and even domains. Finally, our analysis highlights the importance of diverse auxiliary data and optimized prompts for enhanced generalization capacity. Code is available at https://github.com/caoyunkang/AdaCLIP., Comment: Accepted by ECCV 2024

In this paper, we propose a method that, given a partial grid map of an indoor environment built by an autonomous mobile robot, estimates the amount of the explored area represented in the map, as well as whether the uncovered part is still worth being explored or not. Our method is based on a deep convolutional neural network trained on data from partially explored environments with annotations derived from the knowledge of the entire map (which is not available when the network is used for inference). We show how such a network can be used to define a stopping criterion to terminate the exploration process when it is no longer adding relevant details about the environment to the map, saving, on average, 40% of the total exploration time with respect to covering all the area of the environment., Comment: Under review at IEEE RAL

Deep Learning (DL) models have been successfully applied to many applications including biomedical cell segmentation and classification in histological images. These models require large amounts of annotated data which might not always be available, especially in the medical field where annotations are scarce and expensive. To overcome this limitation, we propose a novel pipeline for generating synthetic datasets for cell segmentation. Given only a handful of annotated images, our method generates a large dataset of images which can be used to effectively train DL instance segmentation models. Our solution is designed to generate cells of realistic shapes and placement by allowing experts to incorporate domain knowledge during the generation of the dataset., Comment: 5 pages, Accepted at the International Symposium on Biomedical Imaging (ISBI) 2024, Code available at https://github.com/rb-sl/ExpertDrivenNuclei

Multi-modal Large Language Models (MLLMs) have demonstrated remarkable capabilities in understanding and generating content across various modalities, such as images and text. However, their interpretability remains a challenge, hindering their adoption in critical applications. This research proposes a novel approach to enhance the interpretability of MLLMs by focusing on the image embedding component. We combine an open-world localization model with a MLLM, thus creating a new architecture able to simultaneously produce text and object localization outputs from the same vision embedding. The proposed architecture greatly promotes interpretability, enabling us to design a novel saliency map to explain any output token, to identify model hallucinations, and to assess model biases through semantic adversarial perturbations., Comment: Submitted at BMVC 2024

For more than a decade, deep learning models have been dominating in various 2D imaging tasks. Their application is now extending to 3D imaging, with 3D Convolutional Neural Networks (3D CNNs) being able to process LIDAR, MRI, and CT scans, with significant implications for fields such as autonomous driving and medical imaging. In these critical settings, explaining the model's decisions is fundamental. Despite recent advances in Explainable Artificial Intelligence, however, little effort has been devoted to explaining 3D CNNs, and many works explain these models via inadequate extensions of 2D saliency methods. A fundamental limitation to the development of 3D saliency methods is the lack of a benchmark to quantitatively assess these on 3D data. To address this issue, we propose SE3D: a framework for Saliency method Evaluation in 3D imaging. We propose modifications to ShapeNet, ScanNet, and BraTS datasets, and evaluation metrics to assess saliency methods for 3D CNNs. We evaluate both state-of-the-art saliency methods designed for 3D data and extensions of popular 2D saliency methods to 3D. Our experiments show that 3D saliency methods do not provide explanations of sufficient quality, and that there is margin for future improvements and safer applications of 3D CNNs in critical fields.

Advances in multi-modal embeddings, and in particular CLIP, have recently driven several breakthroughs in Computer Vision (CV). CLIP has shown impressive performance on a variety of tasks, yet, its inherently opaque architecture may hinder the application of models employing CLIP as backbone, especially in fields where trust and model explainability are imperative, such as in the medical domain. Current explanation methodologies for CV models rely on Saliency Maps computed through gradient analysis or input perturbation. However, these Saliency Maps can only be computed to explain classes relevant to the end task, often smaller in scope than the backbone training classes. In the context of models implementing CLIP as their vision backbone, a substantial portion of the information embedded within the learned representations is thus left unexplained. In this work, we propose Concept Visualization (ConVis), a novel saliency methodology that explains the CLIP embedding of an image by exploiting the multi-modal nature of the embeddings. ConVis makes use of lexical information from WordNet to compute task-agnostic Saliency Maps for any concept, not limited to concepts the end model was trained on. We validate our use of WordNet via an out of distribution detection experiment, and test ConVis on an object localization benchmark, showing that Concept Visualizations correctly identify and localize the image's semantic content. Additionally, we perform a user study demonstrating that our methodology can give users insight on the model's functioning., Comment: Accepted for publication at IJCNN 2024

In several problems involving fluid flows, Computational Fluid Dynamics (CFD) provides detailed quantitative information, and often allows the designer to successfully optimize the system, by minimizing a cost function. Sometimes, however, one cannot improve the system with CFD alone, because a suitable cost function is not readily available: one notable example is diagnosis in medicine. The field of interest considered here is rhinology: a correct air flow is key for the functioning of the human nose, yet the notion of a functionally normal nose is not available, and a cost function cannot be written. An alternative and attractive pathway to diagnosis and surgery planning is offered by data-driven methods. In this work, we consider the machine-learning study of nasal pathologies caused by anatomic malformations, with the aim of understanding whether fluid dynamic features, available after a CFD analysis, are more effective than purely geometric features in the training of a neural network for regression. Our experiments are carried out on an extremely simplified anatomic model and a correspondingly simple CFD approach; nevertheless, they demonstrate that flow-based features perform better than geometry-based ones, and allow the training of a neural network with fewer inputs, a crucial advantage in fields like medicine., Comment: Submitted to Flow

Shape matching is a fundamental problem in computer graphics with many applications. Functional maps translate the point-wise shape-matching problem into its functional counterpart and have inspired numerous solutions over the last decade. Nearly all the solutions based on functional maps rely on the eigenfunctions of the Laplace-Beltrami Operator (LB) to describe the functional spaces defined on the surfaces and then convert the functional correspondences into point-wise correspondences. However, this final step is often error-prone and inaccurate in tiny regions and protrusions, where the energy of LB does not uniformly cover the surface. We propose a new functional basis Principal Components of a Dictionary (PCD) to address such intrinsic limitation. PCD constructs an orthonormal basis from the Principal Component Analysis (PCA) of a dictionary of functions defined over the shape. These dictionaries can target specific properties of the final basis, such as achieving an even spreading of energy. Our experimental evaluation compares seven different dictionaries on established benchmarks, showing that PCD is suited to target different shape-matching scenarios, resulting in more accurate point-wise maps than the LB basis when used in the same pipeline. This evidence provides a promising alternative for improving correspondence estimation, confirming the power and flexibility of functional maps., Comment: 22 pages, 12 figures

We introduce Class Distribution Monitoring (CDM), an effective concept-drift detection scheme that monitors the class-conditional distributions of a datastream. In particular, our solution leverages multiple instances of an online and nonparametric change-detection algorithm based on QuantTree. CDM reports a concept drift after detecting a distribution change in any class, thus identifying which classes are affected by the concept drift. This can be precious information for diagnostics and adaptation. Our experiments on synthetic and real-world datastreams show that when the concept drift affects a few classes, CDM outperforms algorithms monitoring the overall data distribution, while achieving similar detection delays when the drift affects all the classes. Moreover, CDM outperforms comparable approaches that monitor the classification error, particularly when the change is not very apparent. Finally, we demonstrate that CDM inherits the properties of the underlying change detector, yielding an effective control over the expected time before a false alarm, or Average Run Length (ARL$_0$)., Comment: 2022 International Joint Conference on Neural Networks (IJCNN)

Deep neural networks require specific layers to process point clouds, as the scattered and irregular location of 3D points prevents the use of conventional convolutional filters. We introduce the composite layer, a flexible and general alternative to the existing convolutional operators that process 3D point clouds. We design our composite layer to extract and compress the spatial information from the 3D coordinates of points and then combine this with the feature vectors. Compared to mainstream point-convolutional layers such as ConvPoint and KPConv, our composite layer guarantees greater flexibility in network design and provides an additional form of regularization. To demonstrate the generality of our composite layers, we define both a convolutional composite layer and an aggregate version that combines spatial information and features in a nonlinear manner, and we use these layers to implement CompositeNets. Our experiments on synthetic and real-world datasets show that, in both classification, segmentation, and anomaly detection, our CompositeNets outperform ConvPoint, which uses the same sequential architecture, and achieve similar results as KPConv, which has a deeper, residual architecture. Moreover, our CompositeNets achieve state-of-the-art performance in anomaly detection on point clouds. Our code is publicly available at \url{https://github.com/sirolf-otrebla/CompositeNet}., Comment: Revised and accepted for publication on Pattern Recognition on 29 April 2024

We address the problem of online change detection in multivariate datastreams, and we introduce QuantTree Exponentially Weighted Moving Average (QT-EWMA), a nonparametric change-detection algorithm that can control the expected time before a false alarm, yielding a desired Average Run Length (ARL$_0$). Controlling false alarms is crucial in many applications and is rarely guaranteed by online change-detection algorithms that can monitor multivariate datastreams without knowing the data distribution. Like many change-detection algorithms, QT-EWMA builds a model of the data distribution, in our case a QuantTree histogram, from a stationary training set. To monitor datastreams even when the training set is extremely small, we propose QT-EWMA-update, which incrementally updates the QuantTree histogram during monitoring, always keeping the ARL$_0$ under control. Our experiments, performed on synthetic and real-world datastreams, demonstrate that QT-EWMA and QT-EWMA-update control the ARL$_0$ and the false alarm rate better than state-of-the-art methods operating in similar conditions, achieving lower or comparable detection delays.

The chips contained in any electronic device are manufactured over circular silicon wafers, which are monitored by inspection machines at different production stages. Inspection machines detect and locate any defect within the wafer and return a Wafer Defect Map (WDM), i.e., a list of the coordinates where defects lie, which can be considered a huge, sparse, and binary image. In normal conditions, wafers exhibit a small number of randomly distributed defects, while defects grouped in specific patterns might indicate known or novel categories of failures in the production line. Needless to say, a primary concern of semiconductor industries is to identify these patterns and intervene as soon as possible to restore normal production conditions. Here we address WDM monitoring as an open-set recognition problem to accurately classify WDM in known categories and promptly detect novel patterns. In particular, we propose a comprehensive pipeline for wafer monitoring based on a Submanifold Sparse Convolutional Network, a deep architecture designed to process sparse data at an arbitrary resolution, which is trained on the known classes. To detect novelties, we define an outlier detector based on a Gaussian Mixture Model fitted on the latent representation of the classifier. Our experiments on a real dataset of WDMs show that directly processing full-resolution WDMs by Submanifold Sparse Convolutions yields superior classification performance on known classes than traditional Convolutional Neural Networks, which require a preliminary binning to reduce the size of the binary images representing WDMs. Moreover, our solution outperforms state-of-the-art open-set recognition solutions in detecting novelties.

Detecting anomalous regions in images is a frequently encountered problem in industrial monitoring. A relevant example is the analysis of tissues and other products that in normal conditions conform to a specific texture, while defects introduce changes in the normal pattern. We address the anomaly detection problem by training a deep autoencoder, and we show that adopting a loss function based on Complex Wavelet Structural Similarity (CW-SSIM) yields superior detection performance on this type of images compared to traditional autoencoder loss functions. Our experiments on well-known anomaly detection benchmarks show that a simple model trained with this loss function can achieve comparable or superior performance to state-of-the-art methods leveraging deeper, larger and more computationally demanding neural networks., Comment: International Conference on Image Analysis and Processing (ICIAP 2021). NVIDIA Prize winner

Artificial intelligence (AI) systems power the world we live in. Deep neural networks (DNNs) are able to solve tasks in an ever-expanding landscape of scenarios, but our eagerness to apply these powerful models leads us to focus on their performance and deprioritises our ability to understand them. Current research in the field of explainable AI tries to bridge this gap by developing various perturbation or gradient-based explanation techniques. For images, these techniques fail to fully capture and convey the semantic information needed to elucidate why the model makes the predictions it does. In this work, we develop a new form of explanation that is radically different in nature from current explanation methods, such as Grad-CAM. Perception visualization provides a visual representation of what the DNN perceives in the input image by depicting what visual patterns the latent representation corresponds to. Visualizations are obtained through a reconstruction model that inverts the encoded features, such that the parameters and predictions of the original models are not modified. Results of our user study demonstrate that humans can better understand and predict the system's decisions when perception visualizations are available, thus easing the debugging and deployment of deep models as trusted systems., Comment: Accepted paper at BMVC 2021 (Proceedings not available yet)

A growing body of work has shown that deep neural networks are susceptible to adversarial examples. These take the form of small perturbations applied to the model's input which lead to incorrect predictions. Unfortunately, most literature focuses on visually imperceivable perturbations to be applied to digital images that often are, by design, impossible to be deployed to physical targets. We present Adversarial Scratches: a novel L0 black-box attack, which takes the form of scratches in images, and which possesses much greater deployability than other state-of-the-art attacks. Adversarial Scratches leverage B\'ezier Curves to reduce the dimension of the search space and possibly constrain the attack to a specific location. We test Adversarial Scratches in several scenarios, including a publicly available API and images of traffic signs. Results show that, often, our attack achieves higher fooling rate than other deployable state-of-the-art methods, while requiring significantly fewer queries and modifying very few pixels., Comment: This work is published at Pattern Recognition (Elsevier). This paper stems from 'Scratch that! An Evolution-based Adversarial Attack against Neural Networks' for which an arXiv preprint is available at arXiv:1912.02316. Further studies led to a complete overhaul of the work, resulting in this paper

We study black-box adversarial attacks for image classifiers in a constrained threat model, where adversaries can only modify a small fraction of pixels in the form of scratches on an image. We show that it is possible for adversaries to generate localized \textit{adversarial scratches} that cover less than $5\%$ of the pixels in an image and achieve targeted success rates of $98.77\%$ and $97.20\%$ on ImageNet and CIFAR-10 trained ResNet-50 models, respectively. We demonstrate that our scratches are effective under diverse shapes, such as straight lines or parabolic B\a'ezier curves, with single or multiple colors. In an extreme condition, in which our scratches are a single color, we obtain a targeted attack success rate of $66\%$ on CIFAR-10 with an order of magnitude fewer queries than comparable attacks. We successfully launch our attack against Microsoft's Cognitive Services Image Captioning API and propose various mitigation strategies.

We address the problem of detecting changes in multivariate datastreams, and we investigate the intrinsic difficulty that change-detection methods have to face when the data dimension scales. In particular, we consider a general approach where changes are detected by comparing the distribution of the log-likelihood of the datastream over different time windows. Despite the fact that this approach constitutes the frame of several change-detection methods, its effectiveness when data dimension scales has never been investigated, which is indeed the goal of our paper. We show that the magnitude of the change can be naturally measured by the symmetric Kullback-Leibler divergence between the pre- and post-change distributions, and that the detectability of a change of a given magnitude worsens when the data dimension increases. This problem, which we refer to as \emph{detectability loss}, is due to the linear relationship between the variance of the log-likelihood and the data dimension. We analytically derive the detectability loss on Gaussian-distributed datastreams, and empirically demonstrate that this problem holds also on real-world datasets and that can be harmful even at low data-dimensions (say, 10).

RF sensor networks are used to localize people indoor without requiring them to wear invasive electronic devices. These wireless mesh networks, formed by low-power radio transceivers, continuously measure the received signal strength (RSS) of the links. Radio Tomographic Imaging (RTI) is a technique that generates 2D images of the change in the electromagnetic field inside the area covered by the radio transceivers to spot the presence and movements of animates (e.g., people, large animals) or large metallic objects (e.g., cars). Here, we present a RTI system for localizing and tracking people outdoors. Differently than in indoor environments where the RSS does not change significantly with time unless people are found in the monitored area, the outdoor RSS signal is time-variant, e.g., due to rainfalls or wind-driven foliage. We present a novel outdoor RTI method that, despite the nonstationary noise introduced in the RSS data by the environment, achieves high localization accuracy and dramatically reduces the energy consumption of the sensing units. Experimental results demonstrate that the system accurately detects and tracks a person in real-time in a large forested area under varying environmental conditions, significantly reducing false positives, localization error and energy consumption compared to state-of-the-art RTI methods.