Semantic segmentation is a fundamental topic in computer vision. Several deep learning methods have been proposed for semantic segmentation with outstanding results. However, these models require a lot of densely annotated images. To address this problem, we propose a new algorithm that uses Hy-perGraph Convolutional Networks for Weakly-supervised Semantic Segmentation (HyperGCN-WSS). Our algorithm constructs spatial and k-Nearest Neighbor (k-NN) graphs from the images in the dataset to generate the hypergraphs. Then, we train a specialized HyperGraph Convolutional Network (HyperGCN) architecture using some weak signals. The outputs of the HyperGCN are denominated pseudo-labels, which are later used to train a DeepLab model for semantic segmentation. HyperGCN-WSS is evaluated on the PASCAL VOC 2012 dataset for semantic segmentation, using scribbles or clicks as weak signals. Our algorithm shows competitive performance against previous methods.

This book provides an introduction to the field of image processing. The text consists of eighteen chapters, covering some of the following topics: image formation; sampling; quantization; color imagery; image transformation techniques; wavelet transform; image segmentation; pattern classification methods; texture representation and analysis methods; fuzzy set theory; content-based image retrieval and image mining; biomedical image analysis and image mining; dynamic scene analysis, moving object detection, and tracking; image compression; and JPEG standards. The book is written in a clear and effective way. It offers comprehensive surveys on JPEG standards, including details on the emerging JPEG2000 standard. A large part of the book is devoted to a discussion of practical applications of image processing. While the authors could combine some chapters and offer problem sections, the book could serve as a useful monography for students, researchers, and engineers who want to deepen and refresh their knowledge of image processing.

Kernel Methods are algorithms that, by replacing the inner product with an appropriate positive definite function, implicitly perform a nonlinear mapping of the input data into a high-dimensional feature space. In this paper, we present a kernel method for clustering inspired by the classical K-Means algorithm in which each cluster is iteratively refined using a one-class Support Vector Machine. Our method, which can be easily implemented, compares favorably with respect to popular clustering algorithms, like K-Means, Neural Gas, and Self-Organizing Maps, on a synthetic data set and three UCI real data benchmarks (IRIS data, Wisconsin breast cancer database, Spam database).

In this paper, the problem of estimating the intrinsic dimension of a data set is investigated. A fractal-based approach using the Grassberger-Procaccia algorithm is proposed. Since the Grassberger-Procaccia algorithm (1983) performs badly on sets of high dimensionality, an empirical procedure that improves the original algorithm has been developed. The procedure has been tested on data sets of known dimensionality and on time series of Santa Fe competition.

We tested a neural model for variable estimation on a benchmark corresponding to a subsystem of the 320 MW power plant located at Piombino, Italy, namely one of the high pressure feedwater lines. The process simulation is based on a physical approach and was validated on real plant data. The experiments concern the estimation of the dynamic behaviour of seven state variables for the last heater in the feedwater line. The chosen neural model is a modified Resource Allocating Network. The Neural State Estimator (NSE) is structured as a MIMO (Multi Input Multi Output) neural net, able to estimate all the state variables at the same time. The NSE was tested on a realistic amount of data, obtaining industrially relevant results, much better than those obtained by classical estimation methods.