A review on the combination of binary classifiers in multiclass problems

Several real problems involve the classification of data into categories or classes. Given a data set containing data whose classes are known, Machine Learning algorithms can be employed for the induction of a classifier able to predict the class of new data from the same domain, performing the desired discrimination. Some learning techniques are originally conceived for the solution of problems with only two classes, also named binary classification problems. However, many problems require the discrimination of examples into more than two categories or classes. This paper presents a survey on the main strategies for the generalization of binary classifiers to problems with more than two classes, known as multiclass classification problems. The focus is on strategies that decompose the original multiclass problem into multiple binary subtasks, whose outputs are combined to obtain the final prediction.     

A prototype for cartographic human body analysis

A cartographic-oriented model uses algebraic map operations to perform spatial analysis of medical data relative to the human body. A prototype system uses 3D visualization techniques to deliver analysis results. A prototype implementation suggests the model might provide the basis for a medical application tool that introduces new information insight.     

Modeling and solving the rooted distance-constrained minimum spanning tree problem

In this paper we discuss models and methods for solving the rooted distance constrained minimum spanning tree problem which is defined as follows: given a graph G=(V,E)$G=(V,E)$ with node set V={0,1,…,n}$V=\{0,1,\dots ,n\}$ and edge set E$E$, two integer weights, a cost _ce$c_e$ and a delay _we$w_e$ associated with each edge e$e$ of E$E$, and a natural (time limit) number H$H$, we wish to find a spanning tree T$T$ of the graph with minimum total cost and such that the unique path from a specified root node, node 0, to any other node has total delay not greater than H$H$. This problem generalizes the well known hop-constrained spanning tree problem and arises in the design of centralized networks with quality of service constraints and also in package shipment with service guarantee constraints. We present three theoretically equivalent modeling approaches, a column generation scheme, a Lagrangian relaxation combined with subgradient optimization procedure, both based on a path formulation of the problem, and a shortest path (compact) reformulation of the problem which views the underlying subproblem as defined in a layered extended graph. We present results for complete graph instances with up to 40 nodes. Our results indicate that the layered graph path reformulation model is still quite good when the arc weights are reasonably small. Lagrangian relaxation combined with subgradient optimization procedure appears to work much better than column generation and seems to be a quite reasonable approach to the problem for large weight, and even small weight, instances.     

Spike train decoding without spike sorting

We propose a novel paradigm for spike train decoding, which avoids entirely spike sorting based on waveform measurements. This paradigm directly uses the spike train collected at recording electrodes from thresholding the bandpassed voltage signal. Our approach is a paradigm, not an algorithm, since it can be used with any of the current decoding algorithms, such as population vector or likelihood-based algorithms. Based on analytical results and an extensive simulation study, we show that our paradigm is comparable to, and sometimes more efficient than, the traditional approach based on well-isolated neurons and that it remains efficient even when all electrodes are severely corrupted by noise, a situation that would render spike sorting particularly difficult. Our paradigm will also save time and computational effort, both of which are crucially important for successful operation of real-time brain-machine interfaces. Indeed, in place of the lengthy spike-sorting task of the traditional approach, it involves an exact expectation EM algorithm that is fast enough that it could also be left to run during decoding to capture potential slow changes in the states of the neurons.     

Brain reading using full brain support vector machines for object recognition: there is no "face" identification area

Over the past decade, object recognition work has confounded voxel response detection with potential voxel class identification. Consequently, the claim that there are areas of the brain that are necessary and sufficient for object identification cannot be resolved with existing associative methods (e.g., the general linear model) that are dominant in brain imaging methods. In order to explore this controversy we trained full brain (40,000 voxels) single TR (repetition time) classifiers on data from 10 subjects in two different recognition tasks on the most controversial classes of stimuli (house and face) and show 97.4% median out-of-sample (unseen TRs) generalization. This performance allowed us to reliably and uniquely assay the classifier's voxel diagnosticity in all individual subjects' brains. In this two-class case, there may be specific areas diagnostic for house stimuli (e.g., LO) or for face stimuli (e.g., STS); however, in contrast to the detection results common in this literature, neither the fusiform face area nor parahippocampal place area is shown to be uniquely diagnostic for faces or places, respectively.     

Model of Frequency Analysis in the Visual Cortex and the Shape from Texture Problem

This paper addresses the question: at the level of cortical cells present in the primary area V1, is the information sufficient to extract the local perspective from the texture? Starting from a model of complex cells in visual area V1, we propose a biologically plausible algorithm for frequency analysis applied to the shape from texture problem. First, specific log-normal filters are designed in replacement of the classical Gabor filters because of their theoretical properties and of their biological plausibility. These filters are separable in frequency and orientation and they better sample the image spectrum which makes them appropriate for any pattern analysis technique. A method to estimate the local frequency in the image, which discards the need to choose the best local scale, is designed. Based on this frequency analysis model, a local decomposition of the image into patches leads to the estimation of the local frequency variation which is used to solve the problem of recovering the shape from the texture. From the analytical relation between the local frequency and the geometrical parameters, under perspective projection, it is possible to recover the orientation and the shape of the original image. The accuracy of the method is evaluated and discussed on different kind of textures, both regular and irregular, with planar and curved surfaces and also on natural scenes and psychophysical stimuli. It compares favorably to the best existing methods, with in addition, a low computational cost. The biological plausibility of the model is finally discussed.     

Detailed Real-Time Urban 3D Reconstruction from Video

The paper presents a system for automatic, geo-registered, real-time 3D reconstruction from video of urban scenes. The system collects video streams, as well as GPS and inertia measurements in order to place the reconstructed models in geo-registered coordinates. It is designed using current state of the art real-time modules for all processing steps. It employs commodity graphics hardware and standard CPU’s to achieve real-time performance. We present the main considerations in designing the system and the steps of the processing pipeline. Our system extends existing algorithms to meet the robustness and variability necessary to operate out of the lab. To account for the large dynamic range of outdoor videos the processing pipeline estimates global camera gain changes in the feature tracking stage and efficiently compensates for these in stereo estimation without impacting the real-time performance. The required accuracy for many applications is achieved with a two-step stereo reconstruction process exploiting the redundancy across frames. We show results on real video sequences comprising hundreds of thousands of frames.     

Multi-Class Segmentation with Relative Location Prior

Multi-class image segmentation has made significant advances in recent years through the combination of local and global features. One important type of global feature is that of inter-class spatial relationships. For example, identifying “tree” pixels indicates that pixels above and to the sides are more likely to be “sky” whereas pixels below are more likely to be “grass.” Incorporating such global information across the entire image and between all classes is a computational challenge as it is image-dependent, and hence, cannot be precomputed. In this work we propose a method for capturing global information from inter-class spatial relationships and encoding it as a local feature. We employ a two-stage classification process to label all image pixels. First, we generate predictions which are used to compute a local relative location feature from learned relative location maps. In the second stage, we combine this with appearance-based features to provide a final segmentation. We compare our results to recent published results on several multi-class image segmentation databases and show that the incorporation of relative location information allows us to significantly outperform the current state-of-the-art.     

Inter-Image Statistics for 3D Environment Modeling

In this article we present a method for automatically recovering complete and dense depth maps of an indoor environment by fusing incomplete data for the 3D environment modeling problem. The geometry of indoor environments is usually extracted by acquiring a huge amount of range data and registering it. By acquiring a small set of intensity images and a very limited amount of range data, the acquisition process is considerably simplified, saving time and energy consumption. In our method, the intensity and partial range data are registered first by using an image-based registration algorithm. Then, the missing geometric structures are inferred using a statistical learning method that integrates and analyzes the statistical relationships between the visual data and the available depth on terms of small patches. Experiments on real-world data on a variety of sampling strategies demonstrate the feasibility of our method. Electronic Supplementary Material  The online version of this article () contains supplementary material, which is available to authorized users.     

Motion Planning for Haptic Guidance

Haptic devices allow a user to feel either reaction forces from virtual interactions or reaction forces reflected from a remote site during a bilateral teleoperation task. Also, guiding forces can be exerted to train the user in the performance of a virtual task or to assist him/her to safely teleoperate a robot. The generation of guiding forces relies on the existence of a motion plan that provides the direction to be followed to reach the goal from any free configuration of the configuration space (-space). This paper proposes a method to obtain such a plan that interleaves a sampling-based exploration of -space with an efficient computation of harmonic functions. A deterministic sampling sequence (with a bias based on harmonic function values) is used to obtain a hierarchical cell decomposition model of -space. A harmonic function is iteratively computed over the partially known model using a novel approach. The harmonic function is the navigation function used as motion plan. The approach has been implemented in a planner (called the Kautham planner) that, given an initial and a goal configuration, provides: (a) a channel of cells connecting the cell that contains the initial configuration with the cell that contains the goal configuration; (b) two harmonic functions over the whole -space, one that guides motions towards the channel and another that guides motions within the channel towards the goal; and (c) a path computed over a roadmap built with the free samples of the channel. The harmonic functions and the solution path are then used to generate the guiding forces for the haptic device. The planning approach is illustrated with examples on 2D and 3D workspaces. This work was partially supported by the CICYT projects DPI2005-00112 and DPI2007-63665.