Christian Olivier

A structural/statistical feature based vector for handwritten character recognition

This paper describes the application of structural features to the statistical recognition of handwritten characters. It has been demonstrated that a complete description of the characters, based on the combination of seven different families of features, can be achieved and that the same general-purpose structural/statistical feature based vector thus defined proves efficient and robust on different categories of handwritten characters such as digits, uppercase letters and graphemes.

A novel color image segmentation method and its application to white blood cell image analysis

According to the fact that the H component in HSI color space contains most of the white blood cell information, and the S component contains the structure information of the white blood cell nucleus, we develop an iterative Otsus approach based on circular histogram for the leukocyte segmentation by taking full advantage of this knowledge. Experimental results show that this method works successfully in the segmentation of color smear microscopic images

Recognition of handwritten words using stochastic models

The paper deals with the global recognition of a small lexicon of words, based on a pseudo segmentation stage introducing anchor points. We avoid the difficult problem of segmentating the word into letters and the complexity involved by such models to build possible letter graphs. We use two structural representations of the word, strokes and graphemes, each of them being analyzed using a Markov model. These simple models are individually optimized by a rigorous choice of the order for fitting the structural properties of the observed data using Akaike information criteria. The conditional probability to have a word model, given the observation sequence, is computed by taking into account the length of the sequence. Results of the study are presented on French cheque images.

Combining structural and statistical features for the recognition of handwritten characters

The authors present a feature vector for the recognition of handwritten characters which combines the strengths of both statistical and structural feature extractors. Thanks to a combination of seven complementary families of features (ranging from pure structural to pure statistical and including both local and global features), a complete description of the characters can be achieved thus providing a wide range of identification clues. The recognition system has been tested on three categories of handwritten characters: handwritten well-segmented digits extracted from the NIST Database, uppercase letters collected from US dead letter envelopes and graphemes generated by a handwritten cursive word segmentation performed on US address word images. We thus demonstrate in this paper that high recognition rates with very low substitution rates can he achieved by means of the same general-purpose structural/statistical feature based vector.

A robust joint source channel coding scheme for image transmission over the ionospheric channel

In this paper, we propose a joint source channel coding (JSCC) scheme to the transmission of fixed images for wireless communication applications. The ionospheric channel which presents some characteristics identical to those found on mobile radio channels, like fading, multipath and Doppler effect is our test channel. As this method based on a wavelet transform, a self-organising map (SOM) vector quantization (VQ) optimally mapped on a QAM digital modulation and an unequal error protection (UEP) strategy, this method is particularly well adapted to low bit-rate applications. The compression process consists in applying a SOM VQ on the discrete wavelet transform coefficients and computing several codebooks depending on the sub-images preserved. An UEP is achieved with a correcting code applied on the most significant data. The JSCC consists of an optimal mapping of the VQ codebook vectors on a high spectral efficiency digital modulation. This feature allows preserving the topological organization of the codebook along the transmission chain while keeping a reduced complexity system. This method applied on grey level images can be used for colour images as well. Several tests of transmission for different images have shown the robustness of this method even for high bit error rate (BER>10^-^2). In order to qualify the quality of the image after transmission, we use a PSNR% (peak signal-to-noise ratio) parameter which is the value of the difference of the PSNR after compression at the transmitter and after reception at the receiver. This parameter clearly shows that 95% of the PSNR is preserved when the BER is less than 10^-^2.

Optimal order of Markov models applied to bankchecks

The aim of this study is to show that the optimal order of Markov Model of cursive words can be rigorously stated in order to fit the structural properties of the observed data using Akaike information criterion. The method has been tested on French Postal check amounts up to order 4. An original structural representation of cursive words based on graphemes is used. The conditional probability to have a word model given an observed sequence of graphemes is computed independently of the length of the sequence. The recognition results obtained confirm the optimal order found using Akaike criterion.

Multi-level Arabic Handwritten Words Recognition

In this paper, we present a strategy of Arabic words recognition by combining two levels which are based on global and analytical approaches according to the topological properties of Arabic handwriting. In the first level (global), we consider the visual indices which can be generated by: diacritics and strokes (denoted tracing) that form the main shapes of the word. Each word is described as a sequence of visual indices which is treated by a “global” classifier based on Hidden Markov Model (HMM). In the second level, the word is segmented into graphemes, then each grapheme is transformed into a HMM observation by a vector quantization phase. An analytical HMM is developed in order to manage the observation sequences. At this level the diacritics are not taken in consideration which allows to reduce the number of estimated character models. Finally we combine the two approaches to decide on the class of an unknown word. In fact, the global model serves as a filter. It produces a set of hypotheses to the analytical model, which in turns, defines and outputs the final decision.

Structural analysis of Arabic handwriting: segmentation and recognition

In this paper, a structural method of recognising Arabic handwritten characters is proposed. The major problem in cursive text recognition is the segmentation into characters or into representative strokes. When we segment the cursive portions of words, we take into account the contextual properties of the Arabic grammar and the junction segments connecting the characters to each other along the writing line. The problem of overlapping characters is resolved with a contour-following algorithm associated with the labelling of the detected contours. In the recognition phase, the characters are gathered into ten families of candidate characters with similar shapes. Then a heterarchical analysis follows that checks the pattern via goal-directed feedback control.

Choice of a 2-D causal autoregressive texture model using information criteria

In the context of parametric modeling for image processing, we derive an estimation method for both the order and the parameters of 2-D causal autoregressive model with different geometries of support. Model parameters are estimated from a lattice representation, i.e. based on reflection coefficients. Lattice parameter estimation algorithms offer advantages compared to the Yule-Walker method: they do not require matrix inversion and their computation are robust and fast. For order selection, information criterion (IC) methods are the most commonly used. Therefore our order selection method is based on the combination of an IC and the prediction errors of models computed from the lattice parameter estimation algorithm. In this paper, we favour two consistent criteria compared to the nonconsistent Akaike criterion: the first criterion is a 2-D extension of Bayesian information criterion; the second criterion, noted φβ, extended here to the 2-D case, is a generalization drawn on Rissanens works. Simulations are provided on synthetic and natural textures with quarter plane support and non-symmetrical half plane support. We validate our results on natural textures using the Kullback divergence. The results show the interest of the combination of 2-DFLRLS algorithm and φβ, criterion to characterize natural textures.

Embedded zerotree runlength wavelet image coding

This paper presents a simple and effective image coding technique: the embedded zerotree runlength wavelet (EZRW) image coding algorithm. The main idea behind EZRW is making use of the correlation among the data output by dominant pass and subordinate pass in the embedded zerotree wavelet (EZW) image encoding process. With this algorithm, the compression ratio is better than the EZW algorithm and comparable with that of the SPIHT algorithm for most test images.