Radwa Fathalla - Academia.edu (original) (raw)

Papers by Radwa Fathalla

Applied Sciences

Compactness and separability of data points are two important properties that contribute to the a... more Compactness and separability of data points are two important properties that contribute to the accuracy of machine learning tasks such as classification and clustering. We propose a framework that enhances the goodness criteria of the two properties by transforming the data points to a subspace in the same feature space, where data points of the same class are most similar to each other. Most related research about feature engineering in the input data points space relies on manually specified transformation functions. In contrast, our work utilizes a fully automated pipeline, in which the transformation function is learnt via an autoencoder for extraction of latent representation and multi-layer perceptron (MLP) regressors for the feature mapping. We tested our framework on both standard small datasets and benchmark-simulated small datasets by taking small fractions of their samples for training. Our framework consistently produced the best results in all semi-supervised clusterin...

Abstract. Words have always been important carriers of information. They convey a lot of aspects ... more Abstract. Words have always been important carriers of information. They convey a lot of aspects about images in which they are embedded. Inspite of the many approaches that have been proposed to separate text appearances from images, very few of them have handled Arabic script. This paper presents a technique to extract Arabic words from a variety of colored images with complex backgrounds. In order to accomplish the task we have chosen the Connected Components (CC) approach. It starts with the breakdown of the RGB image into tiny homogeneous regions using the watershed transform, followed by region merging. The resulting CCs are aggregated into blocks, some of which are the candidate words. Each block is then condensed into a single vector holding the values of it features. The features generally describe the geometrical nature of the Arabic script, including a set of invariant moments. The final decision as to classify the blocks as Arabic words or other was left up to a support ...

Intelligent Data Communication Technologies and Internet of Things, 2019

Gene expression data is used in the prediction of many diseases. Autism spectrum disorder (ASD) i... more Gene expression data is used in the prediction of many diseases. Autism spectrum disorder (ASD) is among those diseases, where information on gene expression for selecting and classifying genes are evaluated. The difficulty of selection and identification of the ASD genes remains a major setback in the gene expression analysis of ASD. The objective of this paper is to develop a classification model for ASD subjects. The paper employs: Deep Belief Network (DBN) based on the Gaussian Restricted Boltzmann machine (GRBM). Restricted Boltzmann machine (RBM) is considered a popular graphical model that constructs a latent representation of raw data fed at its input nodes. The model is based on its learning algorithm, namely, contrastive divergence, and information gain (IG) is used as the criterion for gene selection. Our proposed model proves that it can deal with gene expression values efficiently and achieved improvements over classical classification methods. The results show that tha...

This paper presents a novel approach to the computation of primitive geometrical structures, wher... more This paper presents a novel approach to the computation of primitive geometrical structures, where no prior knowledge about the visual scene is available and a high level of noise is expected. We based our work on the grouping principles of proximity and similarity, of points and preliminary models. The former was realized using Minimum Spanning Trees (MST), on which we apply a stable alignment and goodness of fit criteria. As for the latter, we used spectral clustering of preliminary models. The algorithm can be generalized to various model fitting settings, without tuning of run parameters. Experiments demonstrate the significant improvement in the localization accuracy of models in plane, homography and motion segmentation examples. The efficiency of the algorithm is not dependent on fine tuning of run parameters like most others in the field.

The work presented in this dissertation is a step towards effectively incorporating contextual kn... more The work presented in this dissertation is a step towards effectively incorporating contextual knowledge in the task of semantic segmentation. To date, the use of context has been confined to the genre of the scene with a few exceptions in the field. Research has been directed towards enhancing appearance descriptors. While this is unarguably important, recent studies show that computer vision has reached a near-human level of performance in relying on these descriptors when objects have stable distinctive surface properties and in proper imaging conditions. When these conditions are not met, humans exploit their knowledge about the intrinsic geometric layout of the scene to make local decisions. Computer vision lags behind when it comes to this asset. For this reason, we aim to bridge the gap by presenting algorithms for semantic segmentation of building facades making use of scene topological aspects. We provide a classification scheme to carry out segmentation and recognition sim...

Lecture Notes in Computer Science, 2017

We propose an algorithm that provides a pixel-wise classification of building facades. Building f... more We propose an algorithm that provides a pixel-wise classification of building facades. Building facades provide a rich environment for testing semantic segmentation techniques. They come in a variety of styles affecting appearance and layout. On the other hand, they exhibit a degree of stability in the arrangement of structures across different instances. Furthermore, a single image is often composed of a repetitive architectural pattern. We integrate appearance, layout and repetition cues in a single energy function, that is optimized through the TRW-S algorithm to provide a classification of superpixels. The appearance energy is based on scores of a Random Forrest classifier. The feature space is composed of higher-level vectors encoding distance to structure clusters. Layout priors are obtained from locations and structural adjacencies in training data. In addition, priors result from translational symmetry cues acquired from the scene itself through clustering via the α-expansion graphcut algorithm. We are on par with state-of-the-art. We are able to fine tune classifications at the superpixel level, while most methods model all architectural features with bounding rectangles.

2019 IEEE 15th International Scientific Conference on Informatics

Ninth International Conference on Document Analysis and Recognition (ICDAR 2007) Vol 2, 2007

Page 1. Extraction of Arabic Words from Complex Color Image Radwa Fathalla Yasser El Sonbaty Moha... more Page 1. Extraction of Arabic Words from Complex Color Image Radwa Fathalla Yasser El Sonbaty Mohamed A. Ismail College of Computing College of Computing Faculty of Engineering Arab Academy for Sc. & Tech. Arab Academy for Sc. & Tech. Alexandria University ...

... Radwa Fathalla1, Yasser El Sonbaty1 and Mohamed A. Ismail2 ... In the former type , Lienhart ... more ... Radwa Fathalla1, Yasser El Sonbaty1 and Mohamed A. Ismail2 ... In the former type , Lienhart and Wernickle[2] presented a multiresolution system that consists of multiple feed forward neural networks fed with blocks of the edge orientation maps of the image. ...

Proceedings of the British Machine Vision Conference 2014, 2014

We propose an algorithm that provides a pixel-wise classification of building facades. Building f... more We propose an algorithm that provides a pixel-wise classification of building facades. Building facades provide a rich environment for testing semantic segmentation techniques. They come in a variety of styles that reflect both appearance and layout characteristics. On the other hand, they exhibit a degree of stability in the arrangement of structures across different instances. We integrate appearance and layout cues in a single framework. The most likely label based on appearance is obtained through applying the state-of-the-art deep convolution networks. This is further optimized through Restricted Boltzmann Machines (RBM), applied on vertical and horizontal scanlines of facade models. Learning the probability distributions of the models via the RBMs is utilized in two settings. Firstly, we use them in learning from pre-seen facade samples, in the traditional training sense. Secondly, we learn from the test image at hand, in a way the allows the transfer of visual knowledge of the scene from correctly classified areas to others. Experimentally , we are on par with the reported performance results. However, we do not explicitly specify any hand-engineered features that are architectural scene dependent, nor do we include any dataset specific heuristics/thresholds.

Applied Sciences

Compactness and separability of data points are two important properties that contribute to the a... more Compactness and separability of data points are two important properties that contribute to the accuracy of machine learning tasks such as classification and clustering. We propose a framework that enhances the goodness criteria of the two properties by transforming the data points to a subspace in the same feature space, where data points of the same class are most similar to each other. Most related research about feature engineering in the input data points space relies on manually specified transformation functions. In contrast, our work utilizes a fully automated pipeline, in which the transformation function is learnt via an autoencoder for extraction of latent representation and multi-layer perceptron (MLP) regressors for the feature mapping. We tested our framework on both standard small datasets and benchmark-simulated small datasets by taking small fractions of their samples for training. Our framework consistently produced the best results in all semi-supervised clusterin...

Abstract. Words have always been important carriers of information. They convey a lot of aspects ... more Abstract. Words have always been important carriers of information. They convey a lot of aspects about images in which they are embedded. Inspite of the many approaches that have been proposed to separate text appearances from images, very few of them have handled Arabic script. This paper presents a technique to extract Arabic words from a variety of colored images with complex backgrounds. In order to accomplish the task we have chosen the Connected Components (CC) approach. It starts with the breakdown of the RGB image into tiny homogeneous regions using the watershed transform, followed by region merging. The resulting CCs are aggregated into blocks, some of which are the candidate words. Each block is then condensed into a single vector holding the values of it features. The features generally describe the geometrical nature of the Arabic script, including a set of invariant moments. The final decision as to classify the blocks as Arabic words or other was left up to a support ...

Intelligent Data Communication Technologies and Internet of Things, 2019

Gene expression data is used in the prediction of many diseases. Autism spectrum disorder (ASD) i... more Gene expression data is used in the prediction of many diseases. Autism spectrum disorder (ASD) is among those diseases, where information on gene expression for selecting and classifying genes are evaluated. The difficulty of selection and identification of the ASD genes remains a major setback in the gene expression analysis of ASD. The objective of this paper is to develop a classification model for ASD subjects. The paper employs: Deep Belief Network (DBN) based on the Gaussian Restricted Boltzmann machine (GRBM). Restricted Boltzmann machine (RBM) is considered a popular graphical model that constructs a latent representation of raw data fed at its input nodes. The model is based on its learning algorithm, namely, contrastive divergence, and information gain (IG) is used as the criterion for gene selection. Our proposed model proves that it can deal with gene expression values efficiently and achieved improvements over classical classification methods. The results show that tha...

This paper presents a novel approach to the computation of primitive geometrical structures, wher... more This paper presents a novel approach to the computation of primitive geometrical structures, where no prior knowledge about the visual scene is available and a high level of noise is expected. We based our work on the grouping principles of proximity and similarity, of points and preliminary models. The former was realized using Minimum Spanning Trees (MST), on which we apply a stable alignment and goodness of fit criteria. As for the latter, we used spectral clustering of preliminary models. The algorithm can be generalized to various model fitting settings, without tuning of run parameters. Experiments demonstrate the significant improvement in the localization accuracy of models in plane, homography and motion segmentation examples. The efficiency of the algorithm is not dependent on fine tuning of run parameters like most others in the field.

The work presented in this dissertation is a step towards effectively incorporating contextual kn... more The work presented in this dissertation is a step towards effectively incorporating contextual knowledge in the task of semantic segmentation. To date, the use of context has been confined to the genre of the scene with a few exceptions in the field. Research has been directed towards enhancing appearance descriptors. While this is unarguably important, recent studies show that computer vision has reached a near-human level of performance in relying on these descriptors when objects have stable distinctive surface properties and in proper imaging conditions. When these conditions are not met, humans exploit their knowledge about the intrinsic geometric layout of the scene to make local decisions. Computer vision lags behind when it comes to this asset. For this reason, we aim to bridge the gap by presenting algorithms for semantic segmentation of building facades making use of scene topological aspects. We provide a classification scheme to carry out segmentation and recognition sim...

Lecture Notes in Computer Science, 2017

We propose an algorithm that provides a pixel-wise classification of building facades. Building f... more We propose an algorithm that provides a pixel-wise classification of building facades. Building facades provide a rich environment for testing semantic segmentation techniques. They come in a variety of styles affecting appearance and layout. On the other hand, they exhibit a degree of stability in the arrangement of structures across different instances. Furthermore, a single image is often composed of a repetitive architectural pattern. We integrate appearance, layout and repetition cues in a single energy function, that is optimized through the TRW-S algorithm to provide a classification of superpixels. The appearance energy is based on scores of a Random Forrest classifier. The feature space is composed of higher-level vectors encoding distance to structure clusters. Layout priors are obtained from locations and structural adjacencies in training data. In addition, priors result from translational symmetry cues acquired from the scene itself through clustering via the α-expansion graphcut algorithm. We are on par with state-of-the-art. We are able to fine tune classifications at the superpixel level, while most methods model all architectural features with bounding rectangles.

2019 IEEE 15th International Scientific Conference on Informatics

Ninth International Conference on Document Analysis and Recognition (ICDAR 2007) Vol 2, 2007

Page 1. Extraction of Arabic Words from Complex Color Image Radwa Fathalla Yasser El Sonbaty Moha... more Page 1. Extraction of Arabic Words from Complex Color Image Radwa Fathalla Yasser El Sonbaty Mohamed A. Ismail College of Computing College of Computing Faculty of Engineering Arab Academy for Sc. & Tech. Arab Academy for Sc. & Tech. Alexandria University ...

... Radwa Fathalla1, Yasser El Sonbaty1 and Mohamed A. Ismail2 ... In the former type , Lienhart ... more ... Radwa Fathalla1, Yasser El Sonbaty1 and Mohamed A. Ismail2 ... In the former type , Lienhart and Wernickle[2] presented a multiresolution system that consists of multiple feed forward neural networks fed with blocks of the edge orientation maps of the image. ...

Proceedings of the British Machine Vision Conference 2014, 2014

We propose an algorithm that provides a pixel-wise classification of building facades. Building f... more We propose an algorithm that provides a pixel-wise classification of building facades. Building facades provide a rich environment for testing semantic segmentation techniques. They come in a variety of styles that reflect both appearance and layout characteristics. On the other hand, they exhibit a degree of stability in the arrangement of structures across different instances. We integrate appearance and layout cues in a single framework. The most likely label based on appearance is obtained through applying the state-of-the-art deep convolution networks. This is further optimized through Restricted Boltzmann Machines (RBM), applied on vertical and horizontal scanlines of facade models. Learning the probability distributions of the models via the RBMs is utilized in two settings. Firstly, we use them in learning from pre-seen facade samples, in the traditional training sense. Secondly, we learn from the test image at hand, in a way the allows the transfer of visual knowledge of the scene from correctly classified areas to others. Experimentally , we are on par with the reported performance results. However, we do not explicitly specify any hand-engineered features that are architectural scene dependent, nor do we include any dataset specific heuristics/thresholds.