Open Access. Powered by Scholars. Published by Universities.®
Physical Sciences and Mathematics Commons™
Open Access. Powered by Scholars. Published by Universities.®
- Keyword
-
- Automatic (4)
- Image (4)
- Recognition (4)
- Classification (3)
- Detection (3)
-
- Facial (3)
- Radar (3)
- Adaptive (2)
- Algorithm (2)
- Approach (2)
- Based (2)
- Doppler (2)
- Estimation (2)
- Expression (2)
- Feature (2)
- Hierarchical (2)
- Human (2)
- Improved (2)
- Motion (2)
- Multi (2)
- Neural (2)
- Reduced (2)
- Resolution (2)
- Restoration (2)
- Selection (2)
- Spectrograms (2)
- Training (2)
- Vector (2)
- Visual (2)
- 2 (1)
Articles 1 - 30 of 30
Full-Text Articles in Physical Sciences and Mathematics
Efficient Svm Training With Reduced Weighted Samples, Son Lam Phung, Giang Hoang Nguyen, Abdesselam Bouzerdoum
Efficient Svm Training With Reduced Weighted Samples, Son Lam Phung, Giang Hoang Nguyen, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
This paper presents an efficient training approach for support vector machines that will improve their ability to learn from a large or imbalanced data set. Given an original training set, the proposed approach applies unsupervised learning to extract a smaller set of salient training exemplars, which are represented by weighted cluster centers and the target outputs. In subsequent supervised learning, the objective function is modified by introducing a weight for each new training sample and the corresponding penalty term. In this paper, we investigate two methods of defining the weight based on cluster vectors. The proposed SVM training is implemented …
Fuzzy Logic-Based Image Fusion For Multi-View Through-The-Wall Radar, Cher Hau Seng, Abdesselam Bouzerdoum, Fok Hing Chi Tivive, Moeness G. Amin
Fuzzy Logic-Based Image Fusion For Multi-View Through-The-Wall Radar, Cher Hau Seng, Abdesselam Bouzerdoum, Fok Hing Chi Tivive, Moeness G. Amin
Professor Salim Bouzerdoum
In this paper, we propose a new technique for image fusion in multi-view through-the-wall radar imaging system. As most existing image fusion methods for through-the-wall radar imaging only consider a global fusion operator, it is desirable to consider the differences between each pixel using a local operator. Here, we present a fuzzy logic-based method for pixel-wise image fusion. The performance of the proposed method is evaluated on both simulated and real data from through-the-wall radar imaging system. Experimental results show that the proposed method yields improved performance, compared to existing methods.
Pedestrian Sensing Using Time-Of-Flight Range Camera, Xue Wei, Son Lam Phung, Abdesselam Bouzerdoum
Pedestrian Sensing Using Time-Of-Flight Range Camera, Xue Wei, Son Lam Phung, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
This paper presents a new approach to detect pedestrians using a time-of-flight range camera, for applications in car safety and assistive navigation of the visually impaired. Using 3-D range images not only enables fast and accurate object segmentation and but also provides useful information such as distances to the pedestrians and their probabilities of collision with the user. In the proposed approach, a 3-D range image is first segmented using a modified local variation algorithm. Three state-of-the-art feature extractors (GIST, SIFT, and HOG) are then used to find shape features for each segmented object. Finally, the SVM is applied to …
A Gender Recognition System Using Shunting Inhibitory Convolutional Neural Networks, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
A Gender Recognition System Using Shunting Inhibitory Convolutional Neural Networks, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
In this paper, we employ shunting inhibitory convolutional neural networks to develop an automatic gender recognition system. The system comprises two modules: a face detector and a gender classifier. The human faces are first detected and localized in the input image. Each detected face is then passed to the gender classifier to determine whether it is a male or female. Both the face detection and gender classification modules employ the same neural network architecture; however, the two modules are trained separately to extract different features for face detection and gender classification. Tested on two different databases, Web and BioID database, …
A Pyramidal Neural Network For Visual Pattern Recognition, Son Lam Phung, A. Bouzerdoum
A Pyramidal Neural Network For Visual Pattern Recognition, Son Lam Phung, A. Bouzerdoum
Professor Salim Bouzerdoum
In this paper, we propose a new neural architecture for classification of visual patterns that is motivated by the two concepts of image pyramids and local receptive fields. The new architecture, called pyramidal neural network (PyraNet), has a hierarchical structure with two types of processing layers: Pyramidal layers and one-dimensional (1-D) layers. In the new network, nonlinear two-dimensional (2-D) neurons are trained to perform both image feature extraction and dimensionality reduction. We present and analyze five training methods for PyraNet [gradient descent (GD), gradient descent with momentum, resilient backpropagation (RPROP), Polak-Ribiere conjugate gradient (CG), and Levenberg-Marquadrt (LM)] and two choices …
Automatic Parameter Selection For Feature-Enhanced Radar Image Restoration, Moeness G Amin, Cher Hau Seng, Son Lam Phung, Abdesselam Bouzerdoum
Automatic Parameter Selection For Feature-Enhanced Radar Image Restoration, Moeness G Amin, Cher Hau Seng, Son Lam Phung, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
In this paper, we propose a new technique for optimum parameter selection in non-quadratic radar image restoration. Although both the regularization hyper-parameter and the norm value are influential factors in the characteristics of the formed restoration, most existing optimization methods either require memory intensive computation or prior knowledge of the noise. Here, we present a contrast measure-based method for automated hyper-parameter selection. The proposed method is then extended to optimize the norm value used in non-quadratic image formation and restoration. The proposed method is evaluated on the MSTAR public target database and compared to the GCV method. Experimental results show …
Automatic Recognition Of Smiling And Neutral Facial Expressions, Peiyao Li, S L. Phung, Abdesselam Bouzerdoum, Fok Hing Chi Tivive
Automatic Recognition Of Smiling And Neutral Facial Expressions, Peiyao Li, S L. Phung, Abdesselam Bouzerdoum, Fok Hing Chi Tivive
Professor Salim Bouzerdoum
Facial expression is one way humans convey their emotional states. Accurate recognition of facial expressions via image analysis plays a vital role in perceptual human computer interaction, robotics and online games. This paper focuses on recognising the smiling from the neutral facial expression. We propose a face alignment method to address the localisation error in existing face detection methods. In this paper, smiling and neutral facial expression are differentiated using a novel neural architecture that combines fixed and adaptive non-linear 2-D filters. The fixed filters are used to extract primitive features, whereas the adaptive filters are trained to extract more …
A Digital Pixel Sensor Array With Programmable Dynamic Range, A. Kitchen, A. Bermak, Abdesselam Bouzerdoum
A Digital Pixel Sensor Array With Programmable Dynamic Range, A. Kitchen, A. Bermak, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
This paper presents a digital pixel sensor (DPS) array employing a time domain analogue-to-digital conversion (ADC) technique featuring adaptive dynamic range and programmable pixel response. The digital pixel comprises a photodiode, a voltage comparator, and an 8-bit static memory. The conversion characteristics of the ADC are determined by an array-based digital control circuit, which linearizes the pixel response, and sets the conversion range. The ADC response is adapted to different lighting conditions by setting a single clock frequency. Dynamic range compression was also experimentally demonstrated. This clearly shows the potential of the proposed technique in overcoming the limited dynamic range …
On The Analysis Of Background Subtraction Techniques Using Gaussian Mixture Models, Abdesselam Bouzerdoum, Azeddine Beghdadi, Son Lam Phung, Philippe L. Bouttefroy
On The Analysis Of Background Subtraction Techniques Using Gaussian Mixture Models, Abdesselam Bouzerdoum, Azeddine Beghdadi, Son Lam Phung, Philippe L. Bouttefroy
Professor Salim Bouzerdoum
In this paper, we conduct an investigation into background subtraction techniques using Gaussian Mixture Models (GMM) in the presence of large illumination changes and background variations. We show that the techniques used to date suffer from the trade-off imposed by the use of a common learning rate to update both the mean and variance of the component densities, which leads to a degeneracy of the variance and creates “saturated pixels”. To address this problem, we propose a simple yet effective technique that differentiates between the two learning rates, and imposes a constraint on the variance so as to avoid the …
Multi-Resolution Mean-Shift Algorithm For Vector Quantization, P L. M Bouttefroy, A Bouzerdoum, A Beghdadi, S L. Phung
Multi-Resolution Mean-Shift Algorithm For Vector Quantization, P L. M Bouttefroy, A Bouzerdoum, A Beghdadi, S L. Phung
Professor Salim Bouzerdoum
The generation of stratified codebooks, providing a subset of vectors at different scale levels, has become necessary with the emergence of embedded coder/decoder for scalable image and video formats. We propose an approach based on mean-shift, invoking the multi-resolution framework to generate codebook vectors. Applied to the entire image, mean-shift is slow because it requires each sample to converge to a mode of the distribution. The procedure can be sped up with three simple assumptions: kernel truncation, code attraction and trajectory attraction. Here we propose to apply the mean-shift algorithm to the four image subbands generated by a DWT, namely …
Automatic Human Motion Classification From Doppler Spectrograms, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Moeness G. Amin
Automatic Human Motion Classification From Doppler Spectrograms, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Moeness G. Amin
Professor Salim Bouzerdoum
No abstract provided.
Adaptive Hierarchical Architecture For Visual Recognition, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Son Lam Phung, Khan M. Iftekharuddin
Adaptive Hierarchical Architecture For Visual Recognition, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Son Lam Phung, Khan M. Iftekharuddin
Professor Salim Bouzerdoum
We propose a new hierarchical architecture for visual pattern classification. The new architecture consists of a set of fixed, directional filters and a set of adaptive filters arranged in a cascade structure. The fixed filters are used to extract primitive features such as orientations and edges that are present in a wide range of objects, whereas the adaptive filters can be trained to find complex features that are specific to a given object. Both types of filters are based on the biological mechanism of shunting inhibition. The proposed architecture is applied to two problems: pedestrian detection and car detection. Evaluation …
A Fast Neural-Based Eye Detection System, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
A Fast Neural-Based Eye Detection System, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
This paper presents a fast eye detection system which is based on an artificial neural network known as the shunting inhibitory convolutional neural network, or SICoNNet for short. With its two-dimensional network architecture and the use of convolution operators, the eye detection system processes an entire input image and generates the location map of the detected eyes at the output. The network consists of 479 trainable parameters which are adapted by a modified Levenberg-Marquardt training algorithm in conjunction with a bootstrap procedure. Tested on 180 real images, with 186 faces, the accuracy of the eye detector reaches 96.8% with only …
Automatic Classification Of Gpr Signals, W Shao, A Bouzerdoum, S L. Phung, L Su, B Indraratna, C Rujikiatkamjorn
Automatic Classification Of Gpr Signals, W Shao, A Bouzerdoum, S L. Phung, L Su, B Indraratna, C Rujikiatkamjorn
Professor Salim Bouzerdoum
Ground penetrating radar has been widely used in many areas. However, the processing and interpretation of acquired signals remains a challenging task since it requires experienced users to manage the whole operations. In this paper, we propose an automatic classification system to categorise GPR signals based on magnitude spectrum amplitudes and support vector machines. The system is tested on a real-world GPR data set. The experimental results show that our system can correctly distinguish ground penetrating radar signals reflected by different materials.
Image Quality Assessment Using A Neural Network Approach, Abdesselam Bouzerdoum, A. Havstad, A. Beghdadi
Image Quality Assessment Using A Neural Network Approach, Abdesselam Bouzerdoum, A. Havstad, A. Beghdadi
Professor Salim Bouzerdoum
In this paper, we propose a neural network approach to image quality assessment. In particular, the neural network measures the quality of an image by predicting the mean opinion score (MOS) of human observers, using a set of key features extracted from the original and test images. Experimental results, using 352 JPEG/JPEG2000 compressed images, show that the neural network outputs correlate highly with the MOS scores, and therefore, the neural network can easily serve as a correlate to subjective image quality assessment. Using 10-fold cross-validation, the predicted MOS values have a linear correlation coefficient of 0.9744, a Spearman ranked correlation …
A Car Detection System Based On Hierarchical Visual Features, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
A Car Detection System Based On Hierarchical Visual Features, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
In this paper, we address the problem of detecting and localizing cars in still images. The proposed car detection system is based on a hierarchical feature detector in which the processing units are shunting inhibitory neurons. To reduce the training time and complexity of the network, the shunting inhibitory neurons in the first layer are implemented as directional nonlinear filters, whereas the neurons in the second layer have trainable parameters. A multi-resolution processing scheme is implemented so as to detect cars of different sizes, and to reduce the number of false positives during the detection stage, an adaptive thresholding strategy …
A Human Gait Classification Method Based On Radar Doppler Spectrograms, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Moeness G. Amin
A Human Gait Classification Method Based On Radar Doppler Spectrograms, Fok Hing Chi Tivive, Abdesselam Bouzerdoum, Moeness G. Amin
Professor Salim Bouzerdoum
An image classification technique, which has recently been introduced for visual pattern recognition, is successfully applied for human gait classification based on radar Doppler signatures depicted in the time-frequency domain. The proposed method has three processing stages. The first two stages are designed to extract Doppler features that can effectively characterize humanmotion based on the nature of arm swings, and the third stage performs classification. Three types of arm motion are considered: free-arm swings, one-arm confined swings, and no-arm swings. The last two arm motions can be indicative of a human carrying objects or a person in stressed situations. The …
Wavelet Based Nonlocal-Means Super-Resolution For Video Sequences, H Zheng, A Bouzerdoum, S L. Phung
Wavelet Based Nonlocal-Means Super-Resolution For Video Sequences, H Zheng, A Bouzerdoum, S L. Phung
Professor Salim Bouzerdoum
Video sequence resolution enhancement became a popular research area during the last two decades. Although traditional super-resolution techniques have been successful in dealing with image sequences, many constraints such as global translation between frames, have to be imposed to obtain good performance. In this paper, we present a new wavelet-based nonlocal-means (WNLM) framework to bypass the motion estimation stage. It can handle complex motion changes between frames. Compared with the nonlocal-means (NLM) super-resolution framework, the proposed method provides better result in terms of PSNR and faster processing.
Markov Random Fields For Abnormal Behavior Detection On Highways, Son Lam Phung, Philippe L. Bouttefroy, Abdesselam Bouzerdoum, Azeddine Beghdadi
Markov Random Fields For Abnormal Behavior Detection On Highways, Son Lam Phung, Philippe L. Bouttefroy, Abdesselam Bouzerdoum, Azeddine Beghdadi
Professor Salim Bouzerdoum
This paper introduces a new paradigm for abnormal behavior detection relying on the integration of contextual information in Markov random fields. Contrary to traditional methods, the proposed technique models the local density of object feature vector, therefore leading to simple and elegant criterion for behavior classification. We develop a Gaussian Markov random field mixture catering for multi-modal density and integrating the neighborhood behavior into a local estimate. The convergence of the random field is ensured by online learning through a stochastic clustering algorithm. The system is tested on an extensive dataset (over 2800 vehicles) for behavior modeling. The experimental results …
Fast Digital Optical Flow Estimation Based On Emd, Mickael Quelin, Abdesselam Bouzerdoum, Son Lam Phung
Fast Digital Optical Flow Estimation Based On Emd, Mickael Quelin, Abdesselam Bouzerdoum, Son Lam Phung
Professor Salim Bouzerdoum
This paper presents an optical flow estimation technique based on the so called Elementary Motion Detector (EMD). The aim is to provide a fast but not necessarily very accurate system to be used for specific post processing purposes. This model uses a low complexity algorithm for detecting motion in four directions by identifying specific motion templates. By extending the motion templates to different scales of the input video, an Elementary Velocity Detector (EVD) is created. This one can save computation time by estimating different speeds in parallel. Information from the EVD outputs is then used to generate an estimate of …
Texture Classification Using Convolutional Neural Networks, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Texture Classification Using Convolutional Neural Networks, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
In this paper, we propose a convolutional neural network (CoNN) for texture classification. This network has the ability to perform feature extraction and classification within the same architecture, whilst preserving the two-dimensional spatial structure of the input image. Feature extraction is performed using shunting inhibitory neurons, whereas the final classification decision is performed using sigmoid neurons. Tested on images from the Brodatz texture database, the proposed network achieves similar or better classification performance as some of the most popular texture classification approaches, namely Gabor filters, wavelets, quadratic mirror filters (QMF) and co-occurrence matrix methods. Furthermore, The CoNN classifier outperforms these …
A Particle Swarm Optimization Algorithm Based On Orthogonal Design, Jie Yang, Abdesselam Bouzerdoum, Son Lam Phung
A Particle Swarm Optimization Algorithm Based On Orthogonal Design, Jie Yang, Abdesselam Bouzerdoum, Son Lam Phung
Professor Salim Bouzerdoum
The last decade has witnessed a great interest in using evolutionary algorithms, such as genetic algorithms, evolutionary strategies and particle swarm optimization (PSO), for multivariate optimization. This paper presents a hybrid algorithm for searching a complex domain space, by combining the PSO and orthogonal design. In the standard PSO, each particle focuses only on the error propagated back from the best particle, without “communicating” with other particles. In our approach, this limitation of the standard PSO is overcome by using a novel crossover operator based on orthogonal design. Furthermore, instead of the “generating-and-updating” model in the standard PSO, the elitism …
Reduced Training Of Convolutional Neural Networks For Pedestrian Detection, Giang Hoang Nguyen, Son Lam Phung, Abdesselam Bouzerdoum
Reduced Training Of Convolutional Neural Networks For Pedestrian Detection, Giang Hoang Nguyen, Son Lam Phung, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
Pedestrian detection is a vision task with many practical applications in video surveillance, road safety, autonomous driving and military. However, it is much more difficult compared to the detection of other visual objects, because of the tremendous variations in the inner region as well as the outer shape of the pedestrian pattern. In this paper, we propose a pedestrian detection approach that uses convolutional neural network (CNN) to differentiate pedestrian and non-pedestrian patterns. Among several advantages, the CNN integrates feature extraction and classification into one single, fully adaptive structure. It can extract two-dimensional features at increasing scales, and it is …
A Compressive Sensing Approach To Image Restoration, Matthew Kitchener, Abdesselam Bouzerdoum, Son Lam Phung
A Compressive Sensing Approach To Image Restoration, Matthew Kitchener, Abdesselam Bouzerdoum, Son Lam Phung
Professor Salim Bouzerdoum
In this paper the image restoration problem is solved using a Compressive Sensing approach, and the translation invariant, a Trous, undecimated wavelet transform. The problem is cast as an unconstrained optimization problem which is solved using the Fletcher-Reeves nonlinear conjugate gradient method. A comparison based on experimental results shows that the proposed method achieves comparable if not better performance as other state-of-the-art techniques.
A New Approach To Sparse Image Representation Using Mmv And K-Svd, Jie Yang, Abdesselam Bouzerdoum, Son Lam Phung
A New Approach To Sparse Image Representation Using Mmv And K-Svd, Jie Yang, Abdesselam Bouzerdoum, Son Lam Phung
Professor Salim Bouzerdoum
This paper addresses the problem of image representation based on a sparse decomposition over a learned dictionary. We propose an improved matching pursuit algorithm for Multiple Measurement Vectors (MMV) and an adaptive algorithm for dictionary learning based on multi-Singular Value Decomposition (SVD), and combine them for image representation. Compared with the traditional K-SVD and orthogonal matching pursuit MMV (OMPMMV) methods, the proposed method runs faster and achieves a higher overall reconstruction accuracy.
Improved Facial Expression Recognition With Trainable 2-D Filters And Support Vector Machines, Peiyao Li, Son Lam Phung, Abdesselam Bouzerdoum, Fok Hing Chi Tivive
Improved Facial Expression Recognition With Trainable 2-D Filters And Support Vector Machines, Peiyao Li, Son Lam Phung, Abdesselam Bouzerdoum, Fok Hing Chi Tivive
Professor Salim Bouzerdoum
Facial expression is one way humans convey their emotional states. Accurate recognition of facial expressions is essential in perceptual human-computer interface, robotics and mimetic games. This paper presents a novel approach to facial expression recognition from static images that combines fixed and adaptive 2-D filters in a hierarchical structure. The fixed filters are used to extract primitive features. They are followed by the adaptive filters that are trained to extract more complex facial features. Both types of filters are non-linear and are based on the biological mechanism of shunting inhibition. The features are finally classified by a support vector machine. …
Motion Estimation With Adaptive Regularization And Neighborhood Dependent Constraint, Muhammad Wasim Nawaz, Abdesselam Bouzerdoum, Son Lam Phung
Motion Estimation With Adaptive Regularization And Neighborhood Dependent Constraint, Muhammad Wasim Nawaz, Abdesselam Bouzerdoum, Son Lam Phung
Professor Salim Bouzerdoum
Modern variational motion estimation techniques use total variation regularization along with the L1 norm in constant brightness data term. An algorithm based on such homogeneous regularization is unable to preserve sharp edges and leads to increased estimation errors. A better solution is to modify regularizer along strong intensity variations and occluded areas. In addition, using neighborhood information with data constraint can better identify correspondence between image pairs than using only a pointwise data constraint. In this work, we present a novel motion estimation method that uses neighborhood dependent data constraint to better characterize local image structure. The method also uses …
A Shunting Inhibitory Convolutional Neural Network For Gender Classification, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
A Shunting Inhibitory Convolutional Neural Network For Gender Classification, Fok Hing Chi Tivive, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
Demographic features, such as gender, are very important for human recognition and can be used to enhance social and biometric applications. In this paper, we propose to use a class of convolutional neural networks for gender classification. These networks are built upon the concepts of local receptive field processing and weight sharing, which makes them more tolerant to distortions and variations in two dimensional shapes. Tested on two separate data sets, the proposed networks achieve better classification accuracy than the conventional feedforward multilayer perceptron networks. On the Feret benchmark dataset, the proposed convolutional neural networks achieve a classification rate of …
Feature Selection For Facial Expression Recognition, Abdesselam Bouzerdoum, Son Lam Phung, Fok Hing Chi Tivive, Peiyao Li
Feature Selection For Facial Expression Recognition, Abdesselam Bouzerdoum, Son Lam Phung, Fok Hing Chi Tivive, Peiyao Li
Professor Salim Bouzerdoum
In daily interactions, humans convey their emotions through facial expression and other means. There are several facial expressions that reflect distinctive psychological activities such as happiness, surprise or anger. Accurate recognition of these activities via facial image analysis will play a vital role in natural human-computer interfaces, robotics and mimetic games. This paper focuses on the extraction and selection of salient features for facial expression recognition. We introduce a cascade of fixed filters and trainable non-linear 2-D filters, which are based on the biological mechanism of shunting inhibition. The fixed filters are used to extract primitive features, whereas the adaptive …
Improved Learning In Grid-To-Grid Neural Network Via Clustering, William E. White, Khan M. Iftekharuddin, Abdesselam Bouzerdoum
Improved Learning In Grid-To-Grid Neural Network Via Clustering, William E. White, Khan M. Iftekharuddin, Abdesselam Bouzerdoum
Professor Salim Bouzerdoum
The maze traversal problem involves finding the shortest distance to the goal from any position in a maze. Such maze solving problems have been an interesting challenge in computational intelligence. Previous work has shown that grid-to-grid neural networks such as the cellular simultaneous recurrent neural network (CSRN) can effectively solve simple maze traversing problems better than other iterative algorithms such as the feedforward multi layer perceptron (MLP). In this work, we investigate improved learning for the CSRN maze solving problem by exploiting relevant information about the maze. We cluster parts of the maze using relevant state information and show an …