Learning Deconvolution Network for Semantic Segmentation

Since their first inception more than half a century ago, automatic reading systems have evolved substantially, thereby showing impressive performance on machine-printed text. The recognition of handwriting can, however, still be considered an open research problem due to its substantial variation in appearance. With the introduction of Markovian models to the field, a promising modeling and recognition paradigm was established for automatic offline handwriting recognition. However, so far, no standard procedures for building Markov-model-based recognizers could be established though trends toward unified approaches can be identified. It is therefore the goal of this survey to provide a comprehensive overview of the application of Markov models in the research field of offline handwriting recognition, covering both the widely used hidden Markov models and the less complex Markov-chain or n-gram models. First, we will introduce the typical architecture of a Markov-model-based offline handwriting recognition system and make the reader familiar with the essential theoretical concepts behind Markovian models. Then, we will give a thorough review of the solutions proposed in the literature for the open problems how to apply Markov-model-based approaches to automatic offline handwriting recognition.

/pdf/markov-models-for-offline-handwriting-recognition-a-survey-4ijdjn6mf0.pdf

Markov models for offline handwriting recognition: a survey

We propose a new approach to associate supervised learning-based confidence prediction with the stereo matching problem. First of all, we analyze the characteristics of various confidence measures in the regression forest framework to select effective confidence measures using training data. We then train regression forests again to predict the correctness (confidence) of a match by using selected confidence measures. In addition, we present a confidence-based matching cost modulation scheme based on the predicted correctness for improving the robustness and accuracy of various stereo matching algorithms. We apply the proposed scheme to the semi-global matching algorithm to make it robust under unexpected difficulties that can occur in outdoor environments. We verify the proposed confidence measure selection and cost modulation methods through extensive experimentation with various aspects using KITTI and challenging outdoor datasets.

/pdf/leveraging-stereo-matching-with-learning-based-confidence-3io56lbq5z.pdf

Leveraging stereo matching with learning-based confidence measures

The aim of the paper is to present rough set methods of constructing hierarchical classifiers for approximation of complex concepts. Classifiers are constructed on the basis of experimental data sets and domain knowledge that are mainly represented by concept ontology. Information systems, decision tables and decision rules are basic tools for modeling and constructing such classifiers. The general methodology presented here is applied to approximate spatial complex concepts and spatio-temporal complex concepts defined for (un)structured complex objects, to identify the behavioral patterns of complex objects, and to the automated behavior planning for such objects when the states of objects are represented by spatio-temporal concepts requiring approximation. We describe the results of computer experiments performed on real-life data sets from a vehicular traffic simulator and on medical data concerning the infant respiratory failure.

Hierarchical Classifiers for Complex Spatio-temporal Concepts

Stereo vision is a popular technique to infer depth from two or more images. In this field, confidence measures, typically obtained from the analysis of the cost volume, aim at detecting uncertain disparity assignments. As recently proved, multiple confidence measures combined with hand-crafted features extracted from the cost volume can be used also for other purposes and in particular to improve the overall disparity accuracy leveraging on machine learning techniques. In this paper, starting from the observation that recurrent local patterns occurring in the disparity maps can tell a correct assignment from a wrong one, we follow a completely different methodology to infer a novel confidence measure from scratch. Specifically, leveraging on Convolutional Neural Networks, we pose the confidence formulation as a regression problem by analyzing the disparity map provided by a stereo vision system. Once trained on a subset of the KITTI 2012 dataset with the disparity maps provided by the simple block-matching algorithm, our confidence measure outperforms state-of-the-art with two datasets (KITTI 2015 and Middlebury 2014) as well as with two stereo algorithms. The experimental evaluation reported clearly highlights that our approach is capable to better generalize its behavior in different circumstances with respect to state-of-the-art. Finally, not being based on cost volume analysis, our proposal is also potentially suited for out-of-the-box depth generation devices which usually do not expose the cues required by top-performing approaches.

/pdf/learning-from-scratch-a-confidence-measure-3k2n3q9qoa.pdf

Learning from scratch a confidence measure.

This paper introduces a new multi-lateral filter to fuse low-resolution depth maps with high-resolution images. The goal is to enhance the resolution of Time-of-Flight sensors and, at the same time, reduce the noise level in depth measurements. Our approach is based on the joint bilateral upsampling, extended by a new factor that considers the low reliability of depth measurements along the low-resolution depth map edges. Our experimental results show better performances than alternative depth enhancing data fusion techniques.

/pdf/pixel-weighted-average-strategy-for-depth-sensor-data-fusion-b0nujyaw09.pdf

Pixel weighted average strategy for depth sensor data fusion

A camera system comprises a 3D TOF camera for acquiring a camera-perspective range image of a scene and an image processor for processing the range image. The image processor contains a position and orientation calibration routine implemented therein in hardware and/or software, which position and orientation calibration routine, when executed by the image processor, detects one or more planes within a range image acquired by the 3D TOF camera, selects a reference plane among the at least one or more planes detected and computes position and orientation parameters of the 3D TOF camera with respect to the reference plane, such as, e.g., elevation above the reference plane and/or camera roll angle and/or camera pitch angle.

/pdf/3d-time-of-flight-camera-system-and-position-orientation-4r5h3ec3b9.pdf

3d time-of-flight camera system and position/orientation calibration method therefor

A method of recording 3D images of a scene based on the time-of-flight principle comprises illuminating a scene by emitting light carrying an intensity modulation, imaging the scene onto a pixel array using an optical system, detecting, in each pixel, intensity-modulated light reflected from the scene onto the pixel and determining, for each pixel, a distance value based on the phase of light detected in the pixel. The determination of the distance values comprises a phase-sensitive de-convolution of the scene imaged onto the pixel array such that phase errors induced by light spreading in the optical system are compensated for.

Recording of 3D images of a scene

Hand Gesture Recognition is completed on top-view hand images observed by a Time of Flight(ToF) camera in a car. The work attempts to solve two important problems of touchless interactions inside a car. First, low latency identification of the gestures which are unobtrusive for the driver. Second, reducing the labelled data required to train learning based solutions, this is particularly important because labelling of gesture sequences is expensive and exigent. This work attempts to improve the fast detection of hand-gestures by correcting the probability estimate of a Long Short Term Memory (LSTM) network by pose prediction made by a Convolutional Neural Network(CNN). Weak models for hand gesture classes based on five hand poses are designed to assist in the predictioncorrection scheme. A training procedure to reduce the labelled data required for hand pose classification is also introduced. This method tries to utilise the statistical property of the dataset to identify a good initialization of weights for the CNN, here we demonstrate this using the Principal Component Analysis(PCA) embedding of non-labelled hand pose sequences. While solving a nine class hand gesture problem we demonstrate an accuracy of 89.50% which is better than existing systems. We also show that a PCA embedding based initialization improves the classification performance of the CNN based pose classifier.

[POSTER] A Probabilistic Combination of CNN and RNN Estimates for Hand Gesture Based Interaction in Car

This paper investigates different strategies allowing integrationof contextual information during the feature extractionstage of a cursive handwriting HMM-based recognitionsystem. First we propose to use linear discriminant analysis(LDA) in order to integrate the class information duringfeature set building. Secondly several zoning strategies areused to integrate local contextual information. Finally, aweighting technique is proposed in association with zoningwith the aim of integrating handwriting style. Some experimentswere carried out and the results show the interest ofthe proposed strategies.

/pdf/integration-of-contextual-information-in-handwriting-34abpty732.pdf

Frédéric Grandidier

Papers

Pixel weighted average strategy for depth sensor data fusion

3d time-of-flight camera system and position/orientation calibration method therefor

Recording of 3D images of a scene

[POSTER] A Probabilistic Combination of CNN and RNN Estimates for Hand Gesture Based Interaction in Car

Integration of contextual information in handwriting recognition systems