Publications by Costantino Grana

In this paper we address the task of learning how to segment a particular class of objects, by means of a training set of images and their segmentations. In particular we propose a method to overcome the extremely high training time of a previously proposed solution to this problem, Kernelized Structural Support Vector Machines. We employ a one-class SVM working with joint kernels to robustly learn significant support vectors (representative image-mask pairs) and accordingly weight them to build a suitable energy function for the graph cut framework. We report results obtained on two public datasets and a comparison of training times on different training set sizes.

In this paper we propose to extend the well known graph cut segmentation framework by learning superpixel relations and use them to weight superpixel-to-superpixel edges in a superpixel graph. Adjacent superpixel-pairs are analyzed to build an object boundary model, able to discriminate between superpixel-pairs belonging to the same object or placed on the edge between the foreground object and the background. Several superpixel-pair features are investigated and exploited to build a non-linear SVM to learn object boundary appearance. The adoption of this modified graph cut enhances the performance of a previously proposed segmentation method on two publicly available datasets, reaching state-of-the-art results.

In this paper we propose a multimedia solution for the interactive exploration of illuminated manuscripts. We leveraged on the joint exploitation of content-based image retrieval and relevance feedback to provide an effective mechanism to navigate through the manuscript and add custom knowledge in the form of tags. The similarity retrieval between miniature illustrations is based on covariance descriptors, integrating color, spatial and gradient information. The proposed relevance feedback technique, namely Query Remapping Feature Space Warping, accounts for the user’s opinions by accordingly warping the data points. This is obtained by means of a remapping strategy (from the Riemannian space where covariance matrices lie, referring back to Euclidean space) useful to boost the retrieval performance. Experiments are reported to show the quality of the proposal. Moreover, the complete prototype with user interaction, as already showcased at museums and exhibitions, is presented.

This paper reports a novel approach to deal with the problem of Object and Scene recognition extending the traditional Bag of Words approach in two ways. Firstly, a dataset independent method of summarizing local features, based on multivariate Gaussian descriptors, is employed. Secondly, a recently proposed classification technique, particularly suited for high dimensional feature spaces without any dimensionality reduction step, allows to effectively exploit these features. Experiments are performed on two publicly available datasets and demonstrate the effectiveness of our approach when compared to state-of-the-art methods.

In this paper we propose to integrate the recently introduces ORB descriptors in the currently favored approach for image classification, that is the Bag of Words model. In particular the problem to be solved is to provide a clustering method able to deal with the binary string nature of the ORB descriptors. We suggest to use a k-means like approach, called k-majority, substituting Euclidean distance with Hamming distance and majority selected vector as the new cluster center. Results combining this new approach with other features are provided over the ImageCLEF 2011 dataset.

In this paper, the problem of automatic people removal from digital photographs is addressed. Removing unintended people from a scene can be very useful to focus further steps of image analysis only on the object of interest, A supervised segmentation algorithm is presented and tested in several scenarios.

Several local features have become quite popular for concept detection and search, due to their ability to capture distinctive details. Typically a Bag of Words approach is followed, where a codebook is built by quantizing the local features. In this paper, we propose to represent SIFT local features extracted from an image as a multivariate Gaussian distribution, obtaining a mean vector and a covariance matrix. Differently from common techniques based on the Bag of Words model, our solution does not rely on the construction of a visual vocabulary, thus removing the dependence of the image descriptors on the specific dataset and allowing to immediately retargeting the features to different classification and search problems. Experimental results are conducted on two very different Cultural Heritage image archives, composed of illuminated manuscript miniatures, and architectural elements pictures collected from the web, on which the proposed approach outperforms the Bag of Words technique both in classification and retrieval.

Techniques based on Bag Of Words approach represent images by quantizing local descriptors and summarizing their distribution in a histogram. Dierently, in this paper we describe an image as multivariate Gaussian distribution, estimated over the extracted local descriptors. The estimated distribution is mapped to a high-dimensional descriptor, by concatenating the mean vector and the projection of the covariance matrix on the Euclidean space tangent to the Riemannian manifold. To deal with large scale datasets and high dimensional feature spaces the Stochastic Gradient Descent solver is adopted. The experimental results on Caltech-101 and ImageCLEF2011 show that the method obtains competitive performance with state-of-the art approaches.

The diffusion of powerful mobile devices has posed the basis for new applications implementing on the devices (which are embedded devices) sophisticated computer vision and pattern recognition algorithms. This paper describes the implementation of a complete system for automatic recognition of places localized on a map through the recognition of significant signs by means of the camera of a mobile device (smartphone, tablet, etc.). The paper proposes a novel classification algorithm based on the innovative use of bag-of-words on ORB features. The recognition is achieved using a simple yet effective search scheme which exploits GPS localization to limit the possible matches. This simple solution brings several advantages, such as the speed also on limited-resource devices, the usability also with limited training samples and the easiness of adapting to new training samples and classes. The overall architecture of the system is based on a REST-JSON client-server architecture. The experimental results have been conducted in a real scenario and evaluating the different parameters which influence the performance.

Common techniques represent images by quantizing local descriptors and summarizing their distribution in a histogram. In this paper we propose to employ a parametric description and compare its capabilities to histogram based approaches. We use the multivariate Gaussian distribution, applied over the SIFT descriptors, extracted with dense sampling on a spatial pyramid. Every distribution is converted to a high-dimensional descriptor, by concatenating the mean vector and the projection of the covariance matrix on the Euclidean space tangent to the Riemannian manifold. Experiments on Caltech-101 and ImageCLEF2011 are performed using the Stochastic Gradient Descent solver, which allows to deal with large scale datasets and high dimensional feature spaces.