Claudio Cusano

Image annotation using SVM

The paper describes an innovative image annotation tool for classifying image regions in one of seven classes - sky, skin, vegetation, snow, water, ground, and buildings - or as unknown. This tool could be productively applied in the management of large image and video databases where a considerable volume of images/frames there must be automatically indexed. The annotation is performed by a classification system based on a multi-class Support Vector Machine. Experimental results on a test set of 200 images are reported and discussed.

3D face detection using curvature analysis

Face detection is a crucial preliminary in many applications. Most of the approaches to face detection have focused on the use of two-dimensional images. We present an innovative method that combines a feature-based approach with a holistic one for three-dimensional (3D) face detection. Salient face features, such as the eyes and nose, are detected through an analysis of the curvature of the surface. Each triplet consisting of a candidate nose and two candidate eyes is processed by a PCA-based classifier trained to discriminate between faces and non-faces. The method has been tested, with good results, on some 150 3D faces acquired by a laser range scanner.

Self-Adaptive Image Cropping for Small Displays

We propose a new self-adaptive image cropping algorithm where the processing steps are driven by the classification of the images into semantic classes. The algorithm exploits both visual and semantic information. Visual information is obtained from a visual attention model, while semantic information relates to the automatically assigned image genre and to the detection of face and skin regions.

UMB-DB: A database of partially occluded 3D faces

In this paper we present the UMB-DB 3D face database. The database has been built to test algorithms and systems for 3D face analysis in uncontrolled and challenging scenarios, in particular in those cases where faces are occluded. The database is composed of 1473 pairs of depth and color images of 143 subjects. Each subject has been acquired with different facial expressions, and with the face partially occluded by various objects such as eyeglasses, hats, scarves and hands. The total number of occluded acquisitions is 578. The database, that is freely available for research purposes, could be used for various investigations, some of which are suggested in the paper. For the sake of comparison, we report the results of some of the 3D face detection and recognition algorithms in the state of the art.

Gappy PCA Classification for Occlusion Tolerant 3D Face Detection

This paper presents an innovative approach for the detection of faces in three dimensional scenes. The method is tolerant against partial occlusions produced by the presence of any kind of object. The detection algorithm uses invariant properties of the surfaces to segment salient facial features, namely the eyes and the nose. At least two facial features must be clearly visible in order to perform face detection. Candidate faces are then registered using an ICP (Iterative Correspondent Point) based approach aimed to avoid those samples which belong to the occluding objects. The final face versus non-face discrimination is computed by a Gappy PCA (GPCA) classifier which is able to classify candidate faces using only those regions of the surface which are considered to be non-occluded. The algorithm has been tested using the UND database obtaining 100% of correct detection and only one false alarm. The database has been then processed with an artificial occlusions generator producing realistic acquisitions that emulate unconstrained scenarios. A rate of 89.8% of correct detections shows that 3D data is particularly suited for handling occluding objects. The results have been also verified on a small test set containing real world occlusions obtaining 90.4% of correctly detected faces. The proposed approach can be used to improve the robustness of all those systems requiring a face detection stage in non-controlled scenarios.

Three-Dimensional Occlusion Detection and Restoration of Partially Occluded Faces

This paper presents an innovative three dimensional occlusion detection and restoration strategy for the recognition of three dimensional faces partially occluded by unforeseen, extraneous objects. The detection method considers occlusions as local deformations of the face that correspond to perturbations in a space designed to represent non-occluded faces. Once detected, occlusions represent missing information, or “holes” in the faces. The restoration module exploits the information provided by the non-occluded part of the face to recover the whole face, using an appropriate basis for the space in which non-occluded faces lie. The restoration strategy does not depend on the method used to detect occlusions and can also be applied to restore faces in the presence of noise and missing pixels due to acquisition inaccuracies. The strategy has been experimented on the occluded acquisitions taken from the Bosphorus 3D face database. A method for the generation of real-looking occlusions is also presented. Artificial occlusions, applied to the UND database, allowed for an in-depth analysis of the capabilities of our approach. Experimental results demonstrate the robustness and feasibility of our approach.

Color constancy using CNNs

In this work we describe a Convolutional Neural Network (CNN) to accurately predict the scene illumination. Taking image patches as input, the CNN works in the spatial domain without using hand-crafted features that are employed by most previous methods. The network consists of one convolutional layer with max pooling, one fully connected layer and three output nodes. Within the network structure, feature learning and regression are integrated into one optimization process, which leads to a more effective model for estimating scene illumination. This approach achieves state-of-the-art performance on a standard dataset of RAW images. Preliminary experiments on images with spatially varying illumination demonstrate the stability of the local illuminant estimation ability of our CNN.

AUTOMATIC CLASSIFICATION OF DIGITAL PHOTOGRAPHS BASED ON DECISION FORESTS

Annotating photographs with broad semantic labels can be useful in both image processing and content-based image retrieval. We show here how low-level features can be related to semantic photo categories, such as indoor, outdoor and close-up, using decision forests consisting of trees constructed according to CART methodology. We also show how the results can be improved by introducing a rejection option in the classification process. Experimental results on a test set of 4,500 photographs are reported and discussed.

Detection and Restoration of Occlusions for 3D Face Recognition

This paper presents an innovative restoration strategy which allows for an effective recognition of 3D faces, even when they are partially occluded by unforeseen, extraneous objects such as scarves, hats, glasses, and so on. First, the occluded regions are detected by considering their effects on the projections of the faces in a suitable face space; the non-occluded regions are then used to restore the missing information. Any recognition algorithm can be applied as usual to restored faces. This restoration strategy led to very satisfactory results on a test set of 52 three-dimensional faces presenting various kinds of occlusions

Evaluating color texture descriptors under large variations of controlled lighting conditions.

The recognition of color texture under varying lighting conditions remains an open issue. Several features have been proposed for this purpose, ranging from traditional statistical descriptors to features extracted with neural networks. Still, it is not completely clear under what circumstances a feature performs better than others. In this paper, we report an extensive comparison of old and new texture features, with and without a color normalization step, with a particular focus on how these features are affected by small and large variations in the lighting conditions. The evaluation is performed on a new texture database, which includes 68 samples of raw food acquired under 46 conditions that present single and combined variations of light color, direction, and intensity. The database allows us to systematically investigate the robustness of texture descriptors across large variations of imaging conditions.