In this paper we introduce and experiment with a framework for learning local perceptual distance functions for visual recognition. We learn a distance function foreach training image as a combination of elementary distances between patch-based visual features. We apply these combined local distance functions to the tasks of image retrieval and classification of novel images. On the Caltech 101 object recognition benchmark, we achieve 60.3% mean recognition across classes using 15 training images per class, which is better than the best published performance by Zhang, et al.

Image registration and super-resolution are often treated as distinct processes, to be considered sequentially [1, 3, 7]. Hardie et al. demonstrated that the low-resolution image registration can be

In this paper we estimate quadratic models for natural images using Independent Component Analysis (ICA). The used quadratic functions are a natural extension to linear functions (i.e.

We consider the problem of detecting humans and classifying their pose from a single image. Specifically, our goal is to devise a statistical model that simultaneously answerstwo questions: 1) is there a human in the image?

Large numbers of descriptors and large codebooks are needed for good results and this becomes slow using k-means. We introduce Extremely Randomized Clustering Forests - ensembles of randomly created clustering trees - and show that these provide more accurate results, much faster training and testing and good resistance to background clutter in several state-of-the-art image classification tasks.

The local statistical properties of photographic images, when represented in a multi-scale basis, have been described using Gaussian scale mixtures (GSMs). Here, we use this local description to construct a global field of Gaussian scale mixtures (FoGSM).

Efficient coding models predict that the optimal code for natural images is a population oforiented Gabor receptive fields. These results match response properties of neurons in primary visual cortex, but not those in the retina. Does the retina use an optimal code, and if so, what is it optimized for? Previous theories of retinal coding have assumed that the goal is to encode the maximal amount of information about the sensory signal. However, the image sampled by retinal photoreceptors is degraded both by the optics of the eye and by the photoreceptor noise.

Patch-based appearance models are used in a wide range of computer vision applications. Tolearn such models it has previously been necessary to specify a suitable set of patch sizes and shapes by hand. In the jigsaw model presented here, the shape, size and appearance of patches are learned automatically from the repeated structures in a set of training images. By learning such irregularly shaped'jigsaw pieces', we are able to discover both the shape and the appearance of object parts without supervision. When applied to face images, for example, the learned jigsaw pieces are surprisingly strongly associated with face parts of different shapes and scales such as eyes, noses, eyebrows and cheeks, to name a few. We conclude that learning the shape of the patch not only improves the accuracy of appearance-based part detection but also allows for shape-based part detection. This enables parts of similar appearance but different shapes to be distinguished; forexample, while foreheads and cheeks are both skin colored, they have markedly different shapes.

We address the problem of blind motion deblurring from a single image, caused by a few moving objects. In such situations only part of the image may be blurred, and the scene consists of layers blurred in different degrees. Most of of existing blind deconvolution research concentrates at recovering a single blurring kernel for the entire image. However, in the case of different motions, the blur cannot be modeled with a single kernel, and trying to deconvolve the entire image with the same kernel will cause serious artifacts. Thus, the task of deblurring needs to involve segmentation of the image into regions with different blurs.

Electrical Impedance Tomography (EIT) is a functional imaging method that is being developed for bedside use in critical care medicine. Aiming at improving the chest anatomical resolution of EIT images we developed a fuzzy model based on EIT high temporal resolution and the functional information contained in the pulmonary perfusion and ventilation signals. EIT data from an experimental animal model were collected during normal ventilation and apnea while an injection of hypertonic saline was used as a reference . The fuzzy model was elaborated in three parts: a modeling of the heart, a pulmonary map from ventilation images and, a pulmonary map from perfusion images. Image segmentation was performed using a threshold method and a ventilation/perfusion map was generated. EIT images treated by the fuzzy model were compared with the hypertonic saline injection method and CT-scan images, presenting good results in both qualitative (the image obtained by the model was very similar to that of the CT-scan) and quantitative (the ROC curve provided an area equal to 0.93) point of view. Undoubtedly, these results represent an important step in the EIT images area, since they open the possibility of developing EIT-based bedside clinical methods, which are not available nowadays. These achievements could serve as the base to develop EIT diagnosis system for some life-threatening diseases commonly found in critical care medicine.