A5072297116- Image Retrieval and Classification Techniques
- Handwritten Text Recognition Techniques
- Advanced Database Systems and Queries
- Neural Networks and Applications
- Semantic Web and Ontologies
- Data Management and Algorithms
- Advanced Image and Video Retrieval Techniques
- Natural Language Processing Techniques
- Algorithms and Data Compression
- Cutaneous lymphoproliferative disorders research
- Music and Audio Processing
- Cutaneous Melanoma Detection and Management
- Topic Modeling
- Web Data Mining and Analysis
- Machine Learning and Algorithms
- Genetic and rare skin diseases.
- Image Processing and 3D Reconstruction
- Cancer and Skin Lesions
- Soft tissue tumor case studies
- Text and Document Classification Technologies
- Speech and dialogue systems
- Speech Recognition and Synthesis
- Genetic factors in colorectal cancer
- Data Mining Algorithms and Applications
- Vascular Tumors and Angiosarcomas
Sapienza University of Rome
2014-2025
Policlinico Umberto I
2007-2024
Institute of Molecular Medicine
2017
Santa Maria Nuova Hospital
2014
Alder Hey Children's NHS Foundation Trust
2013-2014
SANE Mental Health Charity
2013
All India Institute of Medical Sciences
2013
Center for Special Minimally Invasive and Robotic Surgery
2013
University of Florence
2002-2011
Leeds General Infirmary
2011
Abstract Motivation: Predicting the secondary structure of a protein (alpha-helix, beta-sheet, coil) is an important step towards elucidating its three-dimensional structure, as well function. Presently, best predictors are based on machine learning approaches, in particular neural network architectures with fixed, and relatively short, input window amino acids, centered at prediction site. Although fixed small avoids overfitting problems, it does not permit capturing variable long-rang...
In this paper, we investigate the capabilities of local feedback multilayered networks, a particular class recurrent in which connections are only allowed from neurons to themselves. class, learning can be accomplished by an algorithm that is both space and time. We describe limits properties these networks give some insights on their use for solving practical problems.
Artificial neural networks have been extensively applied to document analysis and recognition. Most efforts devoted the recognition of isolated handwritten printed characters with widely recognized successful results. However, many other processing tasks, like preprocessing, layout analysis, character segmentation, word recognition, signature verification, effectively faced very promising This paper surveys most significant problems in area offline image processing, where connectionist-based...
Syphilis is characterized by a wide range of variable clinical symptoms; therefore, it often referred to as "The Great Imitator". Here, we report the case 69-year-old hepatitis-C-positive MSM patient, who was admitted our clinic due solitary firm painless erythematous maculopapular lesion with central crater-like crust on upper right thigh that occurred two months prior. The dermoscopy showed an erythematous, copper-colored, oval diffuse monomorphic dotted and glomerular vessels, crust,...
We describe an approach for table location in document images. The documents are described by means of a hierarchical representation that is based on the MXY tree. presence hypothesized searching parallel lines tree page. This hypothesis afterwards verified locating perpendicular or white spaces region included between lines. Lastly, located tables can be merged basis proximity and similarity criteria. use optimization method, relies definition appropriate index, allows us to identify,...
We describe a flexible form-reader system capable of extracting textual information from accounting documents, like invoices and bills service companies. In this kind document, the extraction some fields cannot take place without having detected corresponding instruction fields, which are only constrained to range in given domains. propose modeling document's layout by means attributed relational graphs, turn out be very effective for form registration, as well performing focused search...
Proposes a novel unified approach for integrating explicit knowledge and learning by example in recurrent networks. The is represented automaton rules, which are directly injected into the connections of network. This can be accomplished using technique based on linear programming, instead from random initial weights. Learning conceived as refinement process mainly responsible uncertain information management. We present preliminary results problems automatic speech recognition.< <ETX...
Describes a method for classifying document images on the basis of their physical layout. The layout is described by means hierarchical description, modified X-Y tree, that derived from classical tree segmentation algorithm taking into account cuts along lines in addition to white spaces between blocks. In order reduce problems due noise and skew input image, built top regions extracted commercial OCR package. afterwards coded fixed-size representation takes occurrences patterns representing...
We propose an approach for the word-level indexing of modern printed documents which are difficult to recognize using current OCR engines. By means indexing, it is possible retrieve position words in a document, enabling queries involving proximity terms. Web search engines implement this kind allowing users pages on basis their textual content. Nowadays, digital libraries hold collections digitized that can be retrieved either by browsing document images or relying appropriate metadata...
Much attention has recently been paid to the recognition of graphical objects, such as company logos and trademarks. Recognizing these objects facilitates document classes. Some promising results have achieved by using autoassociator-based artificial neural networks (AANN) in presence homogeneously distributed noise. However, performance drops significantly when dealing with spot-noisy logos, where strips or blobs produce a partial obstruction pictures. We propose new approach for training...
Sperm protein 17 (Sp17) is a highly conserved mammalian whose primary function still poorly understood. Immunohistochemistry (IHC) in the human testis reveals presence of Sp17 some spermatocytes and abundantly spermatids. All spermatogonia, Sertoli cells, Leydig cells appear to be immunonegative for Sp17, whereas interstitial are immunopositive. IHC recognized two distinct populations (immunopositive or not Sp17) ejaculated spermatozoa. Although it will necessary clarify why spermatozoa do...
We describe a top-down approach to the segmentation and representation of documents containing tabular structures. Examples these are invoices technical papers with tables. The is based on an extension X-Y trees, where regions split by means cuts along separators (e.g. lines), in addition white spaces. leaves homogeneous information cutting separators. Adjacency links among tree local relationships between corresponding regions.