Ariane Herbulot

Segmentation of Vectorial Image Features Using Shape Gradients and Information Measures

In this paper, we propose to focus on the segmentation of vectorial features (e.g. vector fields or color intensity) using region-based active contours. We search for a domain that minimizes a criterion based on homogeneity measures of the vectorial features. We choose to evaluate, within each region to be segmented, the average quantity of information carried out by the vectorial features, namely the joint entropy of vector components. We do not make any assumption on the underlying distribution of joint probability density functions of vector components, and so we evaluate the entropy using non parametric probability density functions. A local shape minimizer is then obtained through the evolution of a deformable domain in the direction of the shape gradient.The first contribution of this paper lies in the methodological approach used to differentiate such a criterion. This approach is mainly based on shape optimization tools. The second one is the extension of this method to vectorial data. We apply this segmentation method on color images for the segmentation of color homogeneous regions. We then focus on the segmentation of synthetic vector fields and show interesting results where motion vector fields may be separated using both their length and their direction. Then, optical flow is estimated in real video sequences and segmented using the proposed technique. This leads to promising results for the segmentation of moving video objects.

Shape gradient for image segmentation using information theory

The paper deals with video and image segmentation using region based active contours. We consider the problem of segmentation through the minimization of a new criterion based on information theory. We first propose to derive a general criterion based on the probability density function using the notion of shape gradient. This general derivation is then applied to criteria based on information theory, such as the entropy or the conditional entropy for the segmentation of sequences of images. We present experimental results on grayscale images and color videos showing the accuracy of the proposed method.

Shape gradient for multimodal image segmentation using mutual information

This paper deals with video and image segmentation using region based active contours. We propose to search for an optimal domain with regards to a criterion based on information measures such as entropy of mutual information. We use a general derivation framework based on the notion of shape gradient. This general derivation is applied to criteria based on information theory, such as mutual information for the segmentation of sequences of images. Finally, we present experimental results on color video sequences showing the efficiency of the proposed method for face segmentation.

Shape Gradient for Image and Video Segmentation

In this chapter, we propose to concentrate on the research of an optimal domain with regards to a global criterion including region and boundary functionals. A local shape minimizer is obtained through the evolution of a deformable domain in the direction of the shape gradient. Shape derivation tools, coming from shape optimization theory, allow us to easily differentiate region and boundary functionals. We more particularly focus on region functionals involving region-dependent features that are globally attached to the region. A general framework is proposed and illustrated by many examples involving functions of parametric or non parametric probability density functions (pdfs) of image features. Among these functions, we notably study the minimization of information measures such as the entropy for the segmentation of homogeneous regions or the minimization of the distance between pdfs for tracking or matching regions of interest.

Fast HOG based person detection devoted to a mobile robot with a spherical camera

In this paper, we present a fast Histogram of Oriented Gradients (HOG) based person detector. The detector adopts a cascade of rejectors framework by selecting discriminant features via a new proposed feature selection framework based on Binary Integer Programming. The mathematical programming explicitly formulates an optimization problem to select discriminant features taking detection performance and computation time into account. The learning of the cascade classifier and its detection capability are validated using a proprietary dataset acquired using the Ladybug2 spherical camera and the public INRIA person detection dataset. The final detector achieves a comparable detection performance as Dalal and Triggs [2] detector while achieving on average more than 2.5×-8× speed up depending on the training dataset.

Motion and Appearance Nonparametric Joint Entropy for Video Segmentation

This paper deals with video segmentation based on motion and spatial information. Classically, the motion term is based on a motion compensation error (MCE) between two consecutive frames. Defining a motion-based energy as the integral of a function of the MCE over the object domain implicitly results in making an assumption on the MCE distribution: Gaussian for the square function and, more generally, parametric distributions for functions used in robust estimation. However, these assumptions are not necessarily appropriate. Instead, we propose to define the energy as a function of (an estimation of) the MCE distribution. This function was chosen to be a continuous version of the Ahmad-Lin entropy approximation, the purpose being to be more robust to outliers inherently present in the MCE. Since a motion-only constraint can fail with homogeneous objects, the motion-based energy is enriched with spatial information using a joint entropy formulation. The resulting energy is minimized iteratively using active contours. This approach provides a general framework which consists in defining a statistical energy as a function of a multivariate distribution, independently of the features associated with the object of interest. The link between the energy and the features observed or computed on the video sequence is then made through a nonparametric, kernel-based distribution estimation. It allows for example to keep the same energy definition while using different features or different assumptions on the features.

People Detection with Heterogeneous Features and Explicit Optimization on Computation Time

In this paper we present a novel people detector that employs discrete optimization for feature selection. Specifically, we use binary integer programming to mine heterogeneous features taking both detection performance and computation time explicitly into consideration. The final trained detector exhibits low Miss Rates with significant boost in frame rate. For example, it achieves a 2.6% less Miss Rate at 10-4 FPPW compared to Dalal and Triggs HOG detector with a 9.22x speed improvement.

Deterministic Method for Automatic Visual Grading of Seed Food Products

This paper presents a deterministic method for automatic visual grading, designed to solve the industrial problem of evaluation of seed lots. The sample is thrown in bulk onto a tray placed in a chamber for acquiring color image. An image processing method had been developed to separate and characterize each seed. Shape learning is performed on isolated seeds. The collected information is used for the segmentation. A first step is made based on simple criteria such as regions, edges and normals to the boundary. Then, an active contour with shape prior is performed to improve the results.

Automatic Method for Visual Grading of Seed Food Products

This paper presents an automatic method for visual grading, designed to solve the industrial problem of evaluation of seed lots. The sample is thrown in bulk onto a tray placed in a chamber for acquiring color image. An image processing method had been developed to separate and characterize each seed. The approach adopted for the segmentation step is based on the use of marked point processes and active contour, leading to tackle the problem by a technique of energy minimization.

Automated visual grading of grain kernels by machine vision

This paper presents two automatic methods for visual grading, designed to solve the industrial problem of evaluation of seed lots from the characterization of a representative sample. The sample is thrown in bulk onto a tray placed in a chamber for acquiring color image in a controlled and reproducible manner. Two image processing methods have been developed to separate, and then characterize each seed present in the image. A shape learning is performed on isolated seeds. Collected information is used for the segmentation. The first approach adopted for the segmentation step is based on simple criteria such as regions, edges and normals to the boundary. Marked point processes are used in the second approach, leading to tackle the problem by a technique of energy minimization. In both approaches, an active contour with shape prior is performed to improve the results. A classification is done on shape or color descriptors to evaluate the quality of the sample.