Dalila Cherifi

Effect of Noise, Blur and Motion on Global Appearance Face Recognition based Methods performance

In this work, an objective comparison between some common global appearance face recognition based methods (PCA, FLD, SVD, DCT, DWT and WPD) has been carried out when considering some natural effects that may decrease the performances. In particular, effects such as blur, motion, noise and their combination are taken into account. To evaluate the performances, FEI database containing images corresponding to 200 individuals are used. For each individual, 14 positions have been considered. The quality of face reconnaissance is measured using the wellknown Equal Error Rate (EER) criteria. Interesting results are obtained highlighting the superiority, in some specific contexts, of some of the evaluated methods.

Abnormal tissus extraction in MRI brain medical images

This study is a comparison between two image segmentations methods; the first method is based on normal brains tissue recognition then tumor extraction using thresholding method. The second method is classification based on EM segmentation which is used for both brain recognition and tumor extraction. The goal of these methods is to detect, segment, extract, classify and measure properties of the brain normal and abnormal (tumor) tissues

Importance of eyes and eyebrows for face recognition system

In this paper, we evaluate the effect of removing eyes or eyebrows from face image (no left eyebrow, no right eyebrow, no eyebrows, no left eye, no right eye, no eyes, no left eyebrow and no left eye, no right eyebrow and no right eye, no eyebrows and no eyes) on the performance of face recognition system based on Principal Component Analysis (PCA), Singular Value Decomposition (SVD), Discrete Wavelet Decomposition (DWT), Discrete Cosine Transform (DCT) and application of DWT prior SVD (DWT-SVD). The evaluation is carried on the FEI database using the Recognition Rate (RR) and Equal Error Rate (EER) criteria.

Effect of eyes and eyebrows on face recognition system performance

In this paper, we evaluate the effect of removing eyes or eyebrows from face image (no left eyebrow, no right eyebrow, no eyebrows, no left eye, no right eye, no eyes, no left eyebrow and no left eye, no right eyebrow and no right eye, no eyebrows and no eyes) on the performance of face recognition system based on Principal Component Analysis (PCA), Singular Value Decomposition (SVD), Discrete Wavelet Decomposition (DWT), Discrete Cosine Transform (DCT) and application of DWT prior SVD (DWT-SVD). The evaluation is carried on the FEI database using the Recognition Rate (RR) and Equal Error Rate (EER) criteria.

Combining Improved Euler and Runge-Kutta 4th order for Tractography in Diffusion-Weighted MRI

Abstract Diffusion-Weighted Magnetic Resonance Imaging (DW-MRI) provides information about the local microstructure of the white matter across different voxels; this information can be used to visualize large-scale organization of the brain. Most of previously published diffusion magnetic resonance imaging reconstruction methods are linked to their own track integration method. In this work, we have formulated a general, deterministic tractography algorithm (CIERTE), which is a combination of Improved Euler and Range-Kutta fourth-order algorithm using tracking Error which works with voxel level information about fiber orientations including multiple crossings, and employs a range of stopping criteria as those described in EuDX algorithm and FACT. Our CIERTE tractography algorithm is tested on synthetic and real data, fully evaluated using seven metrics of the tractometer evaluation system and positively compared to state-of-the-art tractography algorithms.

Fusion of face recognition methods at score level

The face recognition problem has been extensively studied by many researchers but accuracy is not satisfactory. This work presents analysis and performance evaluation of global methods (PCA, FLD, DCT, DWT), local methods (SIFT, LBP) and all possible fusions of two methods among them. The fusion is done by consolidating the output of multiple feature extraction algorithms at score levels using four fusion rules which are Mean, Maximum, Minimum, and product [1]. The hybrid methods are tested in terms of recognition rate under different working environment and conditions such as pose variation, illumination and facial expression change, adding effects as blur, motion, noise or combination of those effects. Finally, their robustness is tested by adding some objects as wearing a scarf, putting on a hat, background change, wearing glasses and beard.

3D shape modelling of femur

Statistical shape models have become a widely used tool in computer vision and medical image analysis where they are of considerable interest when studying shape variations in anatomical shapes. The objective of this article is to build a 3D statistical shape modeling for a given data; the implemented process goes through those basic steps, first collect the given data then apply the alignment algorithm based on the ICP (iterative closest point) method which in turn relies on procrustes analysis result as a starting point, next we apply fitting algorithm which is also based on ICP. Finally we obtain the model using PCA (principle component analysis).To achieve this work, we have implemented the above process on the femur model data samples given from the SICAS (Swiss Institute for Computer Assisted Surgery) Medical Image Repository which is used by graphics and vision of Basel Research Group of the Basel University (Switzerland) [1].

Impact of Thatcher effect, Double Illusion and Inversion on face recognition

Thatcher effect or Thatcher illusion is a phenomenon where it becomes difficult to detect local feature changes in an upside down face, despite identical changes being obvious in an upright face. In the Thatcher illusion, in which the eyes and mouth are inverted relative to the rest of the face, looks grotesque when shown upright but not when inverted. Face double illusion is formed by replicating the eyes and mouth below the originals. Inversion can be obtained by rotating the image vertically (upside down) and horizontally, or flip it right and left. So, in this paper, we evaluate the consequence of Thatcher effect, Double Illusion and Inversion (Upside down, Horizontal, Right and Left) on the performance of face recognition system based on Principal Component Analysis (PCA), Singular Value Decomposition (SVD), Discrete Wavelet Decomposition (DWT), Discrete Cosine Transform (DCT), application of DWT prior SVD (DWT-SVD) and application of DWT prior PCA (DWTPCA). The evaluation is carried on the FEI database using the Recognition Rate (RR) criterion.

Subband selection in Wavelet Packet Decomposition for face recognition

In this paper, we evaluated the performance of face recognition based on Wavelet Packet Decomposition (WPD) and Principal Component Analysis (PCA) at second level of decomposition where six wavelet families are employed namely: Daubechies, Haar, Coiflets, Symlets Biorthogonal, and Reverse Biorthogonal. Firstly by taking all of the sixteen subbands obtained after the second level of decomposition and combine them using mean and product rules. Then, each subband is run separately with the purpose of selecting among them the ones that provide lowest Equal Error Rate (EER). After that, subbands with lowest EER are combined together using mean and product rules; aiming for dimensionality reduction of the input image as well as increase the performance of the recognition system.