Juan Carlos Briceño

Biometric identification system by lip shape

Biometrics systems based on lip shape recognition are of great interest, but have received little attention in the scientific literature. This is perhaps due to the belief that they have little discriminative power. However, a careful study shows that the difference between lip outlines is greater than that between shapes at different lip images of the same person. So, biometric identification by lip outline is possible. In this paper the lip outline is obtained from a color face picture: the color image is transformed to the gray scale using the transformation of Chang et al. (1994) and binarized with the Ridler and Calvar threshold. Considering the lip centroid as the origin of coordinates, each pixel lip envelope is parameterized with polar (ordered from -/spl pi/ to +/spl pi/) and Cartesian coordinates (ordered as heights and widths). To asses identity, a multilabeled multiparameter hidden Markov model is used with the polar coordinates and a multilayer neural network is applied to Cartesian coordinates. With a database of 50 people an average classification hit ratio of 96.9% and equal error ratio (EER) of 0.015 are obtained.

Pollen classification based on contour features

Conserving earths biodiversity for future generations is a fundamental global task, where automated recognition of pollen species by means of computer vision represents a highly prioritized issue. This work focuses on analysis and classification stages. The morphological details of the contour are proposed as pollen grains discriminative features. The approach has been developed as a robust pollen identification based on an HMM kernel. A Vector Support Machine was used as classifier. The principal contribution in this work, in terms of the use of the HMM is the gradient optimisation problem implementation in the SVM. 47 tropical honey plant species have been classified achieving a mean of 93.8% ± 1.43 of success.

Hand shape identification on multirange images

Abstract A hand-shape based biometric identification system which is independent of the image spectrum range is proposed here. Two different spectrum ranges; visible and mid-range infrared, were used to validated the architecture, which maintained the accuracy and stability levels between ranges. In particular, three public databases were tested, obtaining accuracies over 99.9% using a 40% hold-out cross-validation approach. Discrete Hidden Markov Models (DHMM) representing each target identification class was trained with angular chain descriptors. A kernel was then extracted from the trained DHMM and applied as a feature extraction method. Finally, supervised Support Vector Machines were used to classify the extracted features.

Biometric Identification Based on Hand-Shape Features Using a HMM Kernel

The present work presents a biometric identification system for hand shape identification. The different contours have been coded based on angular descriptions forming a Markov chain descriptor. Hidden Markov Models (HMM), each representing a target identification class, have been trained with such chains. Features have been calculated from a kernel based on the HMM parameters descriptors. Finally, supervised Support Vector Machines were used to classify parameters from the HMM kernel. Firstly, the system was modelled using 60 users to tune up the HMM and HMM+SVM configuration parameters and finally, the system was checked with all database, 144 users with 10 samples per class. Our experiments have obtained similar results per both cases, showing a scalable, stable and robust system. Our experiments have achieved an upper success rate of 99.92%, using four hand samples per class for training mode, and six hand samples for test mode. This success was found using as features the transformation of 100 points hand shape with our HMM kernel, and as classifier Support Vector Machines with lineal separating functions.

Robust identification of persons by lips contour using shape transformation

In this paper we present a biometric approach, based on lip shape. We have performed an image preprocessing, in order to detect the face of a person image. After this, we have enhanced the lips image using a color transformation, and next we do its detection. The parameterization is based on lips contour points. Those points have been transformed by a Hidden Markov Model (HMM) kernel, using a minimization of Fisher Score. Finally, a one-versus-all multiclass supervised approach based on Support Vector Machines (SVM) is applied as a classifier. A database with 50 users and 10 samples per class has been built. A cross-validation strategy have been applied in our experiments, reaching success rates up to 99.6%, using four lip training samples per class, and evaluating with six lip test samples. This success was found using a shape of 150 points, with 40 states in Hidden Markov Model and a RBF kernel for a supervised approach based on Support Vector Machines.

Pollen grains contour analysis on verification approach

Earths biodiversity has been suffering the effects of human contamination, and as a result there are many species of plants and animals that are dying. Automatic recognition of pollen species by means of computer vision helps to locate specific species and through this identification, study all the diseases and predators which affect this specie, so biologist can improve methods to preserve this species. This work focuses on analysis and classification stages. A classification approach using binarization of pollen grain images, contour and feature extraction to locate the pollen grain objects within the images is being proposed. A Hidden Markov Model classifier was used to classify 17 genders and species from 11 different families of tropical honey bees plants achieving a mean of 98.77% of success.

Angular Contour Parameterization for Signature Identification

This present work presents a parameterization system based on angles from signature edge (2D-shape) for off-line signature identification. We have used three different classifiers, the Nearest Neighbor classifier (K-NN), Neural Networks (NN) and Hidden Markov Models (HMM). Our off-line database has 800 writers with 24 samples per each writer; in total, 19200 images have been used in our experiments. We have got a success rate of 84.64%, applying as classifier Hidden Markov Model, and only used the information from this edge detection method.

Transformation of Hand-Shape Features for a Biometric Identification Approach

The present work presents a biometric identification system for hand shape identification. The different contours have been coded based on angular descriptions forming a Markov chain descriptor. Discrete Hidden Markov Models (DHMM), each representing a target identification class, have been trained with such chains. Features have been calculated from a kernel based on the HMM parameter descriptors. Finally, supervised Support Vector Machines were used to classify parameters from the DHMM kernel. First, the system was modelled using 60 users to tune the DHMM and DHMM_kernel+SVM configuration parameters and finally, the system was checked with the whole database (GPDS database, 144 users with 10 samples per class). Our experiments have obtained similar results in both cases, demonstrating a scalable, stable and robust system. Our experiments have achieved an upper success rate of 99.87% for the GPDS database using three hand samples per class in training mode, and seven hand samples in test mode. Secondly, the authors have verified their algorithms using another independent and public database (the UST database). Our approach has reached 100% and 99.92% success for right and left hand, respectively; showing the robustness and independence of our algorithms. This success was found using as features the transformation of 100 points hand shape with our DHMM kernel, and as classifier Support Vector Machines with linear separating functions, with similar success.

Using fisher kernel on 2D-shape identification

This paper proposes to use the Fisher kernel for planar shape recognition. A synthetic experiment with artificial shapes has been built. The difference among shapes is the number of vertexes, links between vertexes, size and rotation. The 2D-shapes are parameterized with sweeping angles in order to obtain scale and rotation invariance. A Hidden Markov Model is used to obtain the Fisher score which feeds the Support Vector Machine based classifier. Noise has been added to the shapes in order to check the robustness of the system against noise. Hit ratio score over 99%, has been obtained, which shows the ability of the Fisher kernel tool for planar shape recognition.

Video Monitoring of a Mite in Honeybee Cells

This works presents a new proposal towards the development of an intelligent system for automatic detection and monitoring of the mite’s movement, being this mite’s the major pest of the honey bees worldwide which attacks the bee’s pupa during its growing stage. The pupa is an early life stage of some insects undergoing transformation. This stage is presented only in a certain type of insects called holometabolous, which are those that should undergo a complete metamorphosis. We propose a monitoring technique based on background subtraction and a frame by frame monitoring approach, keeping track of the mite’s localization in the space. Based on this proposal, our system has obtained 90.98% of accuracy in the mite’s monitoring stage with a standard deviation of 1.25.