6th International Conference on Image Processing and its Applications, 1997
Abstract We present the application of different indexes based on the modal matching technique to... more Abstract We present the application of different indexes based on the modal matching technique to visual search in an image database. The problem of efficiently retrieving the images similar to a user-defined sketch is addressed. Similarity evaluation for scarcely sampled shapes is outlined, as well as the problems related to modal matching between differently sampled objects. To these aims, four different definitions of suitable similarity indexes are introduced and discussed
Fifth International Conference on Image Processing and its Applications, 1995
ABSTRACT A three-step algorithm is developed to segment oil spills from a marine background on Sy... more ABSTRACT A three-step algorithm is developed to segment oil spills from a marine background on Synthetic Aperture Radar (SAR) data. First, filtering is performed to reduce speckle noise. Then fuzzy clustering is carried out to obtain a preliminary partition of the pixels on the basis of their grey level intensities. A very simple cluster validity criterion is tested to determine the optimal number of clusters present in the data. In order to improve segmentation a final step involves a cluster merging procedure using edge information provided by a Sobel operator. The algorithm has been tested on SEASAT images
Dans cette communication on propose un système pour le contrôle automatique du traffic maritime à... more Dans cette communication on propose un système pour le contrôle automatique du traffic maritime à proximité des ports. Le système utilise des réseaux de neurones pour le traitement de séquences d'image formées par un système radar opérant dans la bande X. ...
QUATORZIEME COLLOQUE GRETSI - JUAN-LES-PINS - DU 13 AU 16 SEPTEMBRE 1993 ATM Transmission of Wave... more QUATORZIEME COLLOQUE GRETSI - JUAN-LES-PINS - DU 13 AU 16 SEPTEMBRE 1993 ATM Transmission of Wavelet Transformed Video Images: the Impact of Cell Losses and Transmission Errors Fabrizio Argenti0', Giuliano Benelli<2), Lorenzo Favalli<2), Stefano Ferraresi'2', ...
Proceedings 2011 IEEE International Conference on Spatial Data Mining and Geographical Knowledge Services, 2011
Ship detection involves two correlated tasks — ship target detection and ship wake detection. Thi... more Ship detection involves two correlated tasks — ship target detection and ship wake detection. This paper proposes a method for automatic ship wake detection from SAR image. A localized Radon transform is processed around the ship target to detect linear features after ship target detection. It can smooth the speckle noise and improve the SNR in Radon transform space. So the false alarms can be reduced. If the extracted linear feature is proved to be a ship wake after testing, the wake line can be reconstructed in the original SAR image and the ship orientation can be estimated without ambiguity.
IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development
This paper outlines a system architecture able to reduce the effects of a devastating seismic eve... more This paper outlines a system architecture able to reduce the effects of a devastating seismic event by providing a rapid and reliable damage detection and estimation of the extent and location of the suffered area. This result has been accomplished by the integration of data access and standardization techniques, image processing tools, GIS technology, analytical modeling and communication tools. A
Urinary Tract Infections (UTIs) are a severe public health problem, accounting for more than eigh... more Urinary Tract Infections (UTIs) are a severe public health problem, accounting for more than eight million visits to health care providers each year. High recurrence rates and increasing antimicrobial resistance among uropathogens threaten to greatly increase the economic burden of these infections. Normally, UTIs are diagnosed by traditional methods, based on cultivation of bacteria on Petri dishes, followed by a visual evaluation by human experts. The need of achieving faster and more accurate results, in order to set a targeted and sudden therapy, motivates the design of an automatic solution in place of the standard procedure. In this paper, we propose an algorithm that combines a “bag–of–words” approach with machine learning techniques to recognize infected plates and provide the automatic classification of the bacterial species. Preliminary experimental results are promising and motivate the introduction of a visual word dictionary with respect to using low level visual features.
Urinary Tract Infections (UTIs) represent a significant health problem, both in hospital and comm... more Urinary Tract Infections (UTIs) represent a significant health problem, both in hospital and community–based settings. Normally, UTIs are diagnosed by traditional methods, based on cultivation of bacteria on Petri dishes, followed by a visual evaluation by human experts. In this paper, we present a fully automated system for the screening, that can provide quick and traceable results of UTIs. Actually, based on image processing techniques and machine learning tools, the recognition of bacteria and the colony count are automatically carried out, yielding accurate results. The proposed system, called AID (Automatic Infections Detector) provides support during the whole analysis process: first digital color images of the Petri dishes are automatically captured, then specific preprocessing and spatial clustering algorithms isolate the colonies from the culture ground, finally an accurate classification of the infection types and their severity is performed. Some important aspects of AID are: reduced time, results repeatability, reduced costs.
IEEE International Conference on Image Processing 2005, 2005
ABSTRACT This paper describes an automatic real-time video surveillance system, capable of autono... more ABSTRACT This paper describes an automatic real-time video surveillance system, capable of autonomously learning and signaling anomalous activities of moving objects. To obtain these capabilities, an improved version of the altruistic vector quantization algorithm (AVQ) is proposed. The modified AVQ automatically evaluates the number of trajectory prototypes, and improves the representativeness of the prototypes themselves, so the visual events can be easily and accurately classified. Anomalous behaviors are detected if visual trajectories deviate from the self-learned representations of &quot;typical&quot; behaviors. The system has been implemented by means of standard PCs and TV cameras, and has been tested in many real outdoor contexts in different conditions (night and day). Currently it is used to monitor the storage areas of British Airways at the airport of Peretola (Florence, Italy), and some access gates of Autostrade per FItalia S.p.A. (the main Italian highways company). If the camera field-of-view is changed, the system automatically re-learns new &quot;typical&quot; behaviors and accurately detects anomalous events.
Deep learning is widely applied in bioinformatics and biomedical imaging, due to its ability to p... more Deep learning is widely applied in bioinformatics and biomedical imaging, due to its ability to perform various clinical tasks automatically and accurately. In particular, the application of deep learning techniques for the automatic identification of glomeruli in histopathological kidney images can play a fundamental role, offering a valid decision support system tool for the automatic evaluation of the Karpinski metric. This will help clinicians in detecting the presence of sclerotic glomeruli in order to decide whether the kidney is transplantable or not. In this work, we implemented a deep learning framework to identify and segment sclerotic and non-sclerotic glomeruli from scanned Whole Slide Images (WSIs) of human kidney biopsies. The experiments were conducted on a new dataset collected by both the Siena and Trieste hospitals. The images were segmented using the DeepLab V2 model, with a pre-trained ResNet101 encoder, applied to 512 × 512 patches extracted from the original WS...
2016 7th IEEE International Conference on Cognitive Infocommunications (CogInfoCom), 2016
This paper presents a framework for the monitoring of hospitalized people, including fall detecti... more This paper presents a framework for the monitoring of hospitalized people, including fall detection capabilities, using an environmentally mounted depth imaging sensor. The purpose is to characterize the fall event, depending on the location of the person when the fall event happens. In particular, we distinguish two basic starting point conditions: fall from standing position (e.g. due to blood pressure failure) and fall out of bed (e.g. due to agitation). To achieve this goal, we exploit the context information to adaptively extract the person's silhouette and then reliably tracking the trajectory. If a fall occurs, the system is capable of recognize this event on the basis of the inferred starting condition. The current implementation has been tested on available online datasets and on a self-made dedicated dataset. In this latter dataset, we have included falls from standing position and falls out of bed, even in presence of occlusions.
In this paper, we use Generative Adversarial Networks (GANs) to synthesize high-quality retinal i... more In this paper, we use Generative Adversarial Networks (GANs) to synthesize high-quality retinal images along with the corresponding semantic label-maps, instead of real images during training of a segmentation network. Different from other previous proposals, we employ a two-step approach: first, a progressively growing GAN is trained to generate the semantic label-maps, which describes the blood vessel structure (i.e., the vasculature); second, an image-to-image translation approach is used to obtain realistic retinal images from the generated vasculature. The adoption of a two-stage process simplifies the generation task, so that the network training requires fewer images with consequent lower memory usage. Moreover, learning is effective, and with only a handful of training samples, our approach generates realistic high-resolution images, which can be successfully used to enlarge small available datasets. Comparable results were obtained by employing only synthetic images in plac...
6th International Conference on Image Processing and its Applications, 1997
Abstract We present the application of different indexes based on the modal matching technique to... more Abstract We present the application of different indexes based on the modal matching technique to visual search in an image database. The problem of efficiently retrieving the images similar to a user-defined sketch is addressed. Similarity evaluation for scarcely sampled shapes is outlined, as well as the problems related to modal matching between differently sampled objects. To these aims, four different definitions of suitable similarity indexes are introduced and discussed
Fifth International Conference on Image Processing and its Applications, 1995
ABSTRACT A three-step algorithm is developed to segment oil spills from a marine background on Sy... more ABSTRACT A three-step algorithm is developed to segment oil spills from a marine background on Synthetic Aperture Radar (SAR) data. First, filtering is performed to reduce speckle noise. Then fuzzy clustering is carried out to obtain a preliminary partition of the pixels on the basis of their grey level intensities. A very simple cluster validity criterion is tested to determine the optimal number of clusters present in the data. In order to improve segmentation a final step involves a cluster merging procedure using edge information provided by a Sobel operator. The algorithm has been tested on SEASAT images
Dans cette communication on propose un système pour le contrôle automatique du traffic maritime à... more Dans cette communication on propose un système pour le contrôle automatique du traffic maritime à proximité des ports. Le système utilise des réseaux de neurones pour le traitement de séquences d'image formées par un système radar opérant dans la bande X. ...
QUATORZIEME COLLOQUE GRETSI - JUAN-LES-PINS - DU 13 AU 16 SEPTEMBRE 1993 ATM Transmission of Wave... more QUATORZIEME COLLOQUE GRETSI - JUAN-LES-PINS - DU 13 AU 16 SEPTEMBRE 1993 ATM Transmission of Wavelet Transformed Video Images: the Impact of Cell Losses and Transmission Errors Fabrizio Argenti0', Giuliano Benelli<2), Lorenzo Favalli<2), Stefano Ferraresi'2', ...
Proceedings 2011 IEEE International Conference on Spatial Data Mining and Geographical Knowledge Services, 2011
Ship detection involves two correlated tasks — ship target detection and ship wake detection. Thi... more Ship detection involves two correlated tasks — ship target detection and ship wake detection. This paper proposes a method for automatic ship wake detection from SAR image. A localized Radon transform is processed around the ship target to detect linear features after ship target detection. It can smooth the speckle noise and improve the SNR in Radon transform space. So the false alarms can be reduced. If the extracted linear feature is proved to be a ship wake after testing, the wake line can be reconstructed in the original SAR image and the ship orientation can be estimated without ambiguity.
IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development
This paper outlines a system architecture able to reduce the effects of a devastating seismic eve... more This paper outlines a system architecture able to reduce the effects of a devastating seismic event by providing a rapid and reliable damage detection and estimation of the extent and location of the suffered area. This result has been accomplished by the integration of data access and standardization techniques, image processing tools, GIS technology, analytical modeling and communication tools. A
Urinary Tract Infections (UTIs) are a severe public health problem, accounting for more than eigh... more Urinary Tract Infections (UTIs) are a severe public health problem, accounting for more than eight million visits to health care providers each year. High recurrence rates and increasing antimicrobial resistance among uropathogens threaten to greatly increase the economic burden of these infections. Normally, UTIs are diagnosed by traditional methods, based on cultivation of bacteria on Petri dishes, followed by a visual evaluation by human experts. The need of achieving faster and more accurate results, in order to set a targeted and sudden therapy, motivates the design of an automatic solution in place of the standard procedure. In this paper, we propose an algorithm that combines a “bag–of–words” approach with machine learning techniques to recognize infected plates and provide the automatic classification of the bacterial species. Preliminary experimental results are promising and motivate the introduction of a visual word dictionary with respect to using low level visual features.
Urinary Tract Infections (UTIs) represent a significant health problem, both in hospital and comm... more Urinary Tract Infections (UTIs) represent a significant health problem, both in hospital and community–based settings. Normally, UTIs are diagnosed by traditional methods, based on cultivation of bacteria on Petri dishes, followed by a visual evaluation by human experts. In this paper, we present a fully automated system for the screening, that can provide quick and traceable results of UTIs. Actually, based on image processing techniques and machine learning tools, the recognition of bacteria and the colony count are automatically carried out, yielding accurate results. The proposed system, called AID (Automatic Infections Detector) provides support during the whole analysis process: first digital color images of the Petri dishes are automatically captured, then specific preprocessing and spatial clustering algorithms isolate the colonies from the culture ground, finally an accurate classification of the infection types and their severity is performed. Some important aspects of AID are: reduced time, results repeatability, reduced costs.
IEEE International Conference on Image Processing 2005, 2005
ABSTRACT This paper describes an automatic real-time video surveillance system, capable of autono... more ABSTRACT This paper describes an automatic real-time video surveillance system, capable of autonomously learning and signaling anomalous activities of moving objects. To obtain these capabilities, an improved version of the altruistic vector quantization algorithm (AVQ) is proposed. The modified AVQ automatically evaluates the number of trajectory prototypes, and improves the representativeness of the prototypes themselves, so the visual events can be easily and accurately classified. Anomalous behaviors are detected if visual trajectories deviate from the self-learned representations of &quot;typical&quot; behaviors. The system has been implemented by means of standard PCs and TV cameras, and has been tested in many real outdoor contexts in different conditions (night and day). Currently it is used to monitor the storage areas of British Airways at the airport of Peretola (Florence, Italy), and some access gates of Autostrade per FItalia S.p.A. (the main Italian highways company). If the camera field-of-view is changed, the system automatically re-learns new &quot;typical&quot; behaviors and accurately detects anomalous events.
Deep learning is widely applied in bioinformatics and biomedical imaging, due to its ability to p... more Deep learning is widely applied in bioinformatics and biomedical imaging, due to its ability to perform various clinical tasks automatically and accurately. In particular, the application of deep learning techniques for the automatic identification of glomeruli in histopathological kidney images can play a fundamental role, offering a valid decision support system tool for the automatic evaluation of the Karpinski metric. This will help clinicians in detecting the presence of sclerotic glomeruli in order to decide whether the kidney is transplantable or not. In this work, we implemented a deep learning framework to identify and segment sclerotic and non-sclerotic glomeruli from scanned Whole Slide Images (WSIs) of human kidney biopsies. The experiments were conducted on a new dataset collected by both the Siena and Trieste hospitals. The images were segmented using the DeepLab V2 model, with a pre-trained ResNet101 encoder, applied to 512 × 512 patches extracted from the original WS...
2016 7th IEEE International Conference on Cognitive Infocommunications (CogInfoCom), 2016
This paper presents a framework for the monitoring of hospitalized people, including fall detecti... more This paper presents a framework for the monitoring of hospitalized people, including fall detection capabilities, using an environmentally mounted depth imaging sensor. The purpose is to characterize the fall event, depending on the location of the person when the fall event happens. In particular, we distinguish two basic starting point conditions: fall from standing position (e.g. due to blood pressure failure) and fall out of bed (e.g. due to agitation). To achieve this goal, we exploit the context information to adaptively extract the person's silhouette and then reliably tracking the trajectory. If a fall occurs, the system is capable of recognize this event on the basis of the inferred starting condition. The current implementation has been tested on available online datasets and on a self-made dedicated dataset. In this latter dataset, we have included falls from standing position and falls out of bed, even in presence of occlusions.
In this paper, we use Generative Adversarial Networks (GANs) to synthesize high-quality retinal i... more In this paper, we use Generative Adversarial Networks (GANs) to synthesize high-quality retinal images along with the corresponding semantic label-maps, instead of real images during training of a segmentation network. Different from other previous proposals, we employ a two-step approach: first, a progressively growing GAN is trained to generate the semantic label-maps, which describes the blood vessel structure (i.e., the vasculature); second, an image-to-image translation approach is used to obtain realistic retinal images from the generated vasculature. The adoption of a two-stage process simplifies the generation task, so that the network training requires fewer images with consequent lower memory usage. Moreover, learning is effective, and with only a handful of training samples, our approach generates realistic high-resolution images, which can be successfully used to enlarge small available datasets. Comparable results were obtained by employing only synthetic images in plac...
Uploads
Papers by Alessandro Mecocci