Leizer Schnitman

Highway Traffic Congestion Classification using Holistic Properties

This work proposes a holistic method for highway traffic video classification based on vehicle crowd properties. The method classifies the traffic congestion into three classes: light, medium and heavy. This is done by usage of average crowd density and crowd speed. Firstly, the crowd density is estimated by background subtraction and the crowd speed is performed by pyramidal Kanade-Lucas-Tomasi (KLT) tracker algorithm. The features classification with neural networks show 94.50% of accuracy on experimental results from 254 highway traffic videos of UCSD data set.

Pattern Recognition for Downhole Dynamometer Card in Oil Rod Pump System using Artificial Neural Networks

This paper presents the development of an Artificial Neural Network system for Dynamometer Card pattern recognition in oil well rod pump systems. It covers the establishment of pattern classes and a set of standards for training and validation, the study of descriptors which allow the design and the implementation of features extractor, training, analysis and finally the validation and performance test with a real data base.

Multi-Scale Spectral Residual Analysis to Speed up Image Object Detection

Accuracy in image object detection has been usually achieved at the expense of much computational load. Therefore a trade-off between detection performance and fast execution commonly represents the ultimate goal of an object detector in real life applications. In this present work, we propose a novel method toward that goal. The proposed method was grounded on a multi-scale spectral residual (MSR) analysis for saliency detection. Compared to a regular sliding window search over the images, in our experiments, MSR was able to reduce by 75% (in average) the number of windows to be evaluated by an object detector. The proposed method was thoroughly evaluated over a subset of Label Me dataset (person images), improving detection performance in most cases.

Learning to segment roads for traffic analysis in urban images

Road segmentation plays an important role in many computer vision applications, either for in-vehicle perception or traffic surveillance. In camera-equipped vehicles, road detection methods are being developed for advanced driver assistance, lane departure, and aerial incident detection, just to cite a few. In traffic surveillance, segmenting road information brings special benefits: to automatically wrap regions of traffic analysis (consequently, speeding up flow analysis in videos), to help with the detection of driving violations (to improve contextual information in videos of traffic), and so forth. Methods and techniques can be used interchangeably for both types of application. Particularly, we are interested in segmenting road regions from the remaining of an image, aiming to support traffic flow analysis tasks. In our proposed method, road segmentation relies on a superpixel detection based on a novel edge density estimation method; in each superpixel, priors are extracted from features of gray-amount, texture homogeneity, traffic motion and horizon line. A feature vector with all those priors feeds a support vector machine classifier, which ultimately takes the superpixel-wise decision of being a road or not. A dataset of challenging scenes was gathered from traffic video surveillance cameras, in our city, to demonstrate the effectiveness of the method.

Analysis of exact linearization and aproximate feedback linearization techniques

This paper presents a study of linear control systems based on exact feedback linearization and approximate feedback linearization. As exact feedback linearization is applied, a linear controller can perform the control objectives. The approximate feedback linearization is required when a nonlinear system presents a noninvolutive property. It uses a Taylor series expansion in order to compute a nonlinear transformation of coordinates to satisfy the involutivity conditions.

Exact linearization and fuzzy logic applied to the control of a Magnetic Levitation System

In recent years the area of nonlinear control systems has been the subject of many studies. Computational developments have enabled more complex applications to provide solutions to nonlinear problems. The purpose of this paper is to use a combination of two techniques to control a nonlinear system: the Magnetic Levitation System. First, the exact linearization technique with state feedback is applied to obtain a linear system. Second, the linearization is made via direct cancellation of nonlinear functions, which represent the phenomenological model of the system. Finally, to deal with the presence of uncertainty in the system model, an adaptive controller is used. The controller is based on fuzzy logic to estimate the functions that contain the nonlinearities of the system. The fuzzy system is a zero-order Takagi-Sugeno-Kang structure and the adaptive controller is implemented in a simulated environment (Matlab Simulink ©). The methodology guarantees the convergence of the estimates to their optimal values, and in turn the overall stability of the system. The results show the controller output signal tracks a reference input signal. For future work this adaptive controller should be implemented in a real physical system.

An architecture of sensor fusion for spatial location of objects in mobile robotics

Each part of a mobile robot has particular aspects of its own, which must be integrated in order to successfully conclude a specific task. Among these parts, sensing enables to construct a representation of landmarks of the surroundings with the goal of supplying relevant information for the robot’s navigation. The present work describes the architecture of a perception system based on data fusion from a CMOS camera and distance sensors. The aim of the proposed architecture is the spatial location of objects on a soccer field. An SVM is used for both recognition and object location and the process of fusion is made by means of a fuzzy system, using a TSK model.

A hybrid neural architecture for fuzzy rules extraction

This article proposes a hybrid neural architecture (i.e. neuro-fuzzy) capable to generate Takagi-Sugeno fuzzy rules and to adjust the membership functions and the output functions. The main idea is to apply this algorithm to function approach tasks, whose input-output relationships are the only available information. The proposed algorithm spares previous data analysis. Initially, the user only defines the number of fuzzy rules that the system should produce. This model has the convenience of not requesting the previous and empiric knowledge of the fuzzy rules structure. The proposed model is initiated with an architecture fully connected; in other words, implicit rules do not exist in the connections of the net. Instead of this, the connections between the entrance layer and the T-norms are linked by synaptic weights. These weights must be adjusted and, in a posterior stage, are partially eliminated. The training process occurs in four stages: in the first stage is adjusted the synapses and the function parameters. The second stage occurs when the quadratic error reaches acceptable values. Then, a part of the synapses introduced between the first and the second layer are eliminated. In this stage the proposed net is equaled to the ANFIS model, because the disabled synapses reveal the base rules generated automatically by the algorithm. In the third stage, the adjustment process continues from an identical way to the ANFIS. Finally, in the fourth stage, is eliminated the lower relevant rules. This process makes the base of rules more intelligible.

Constraining image object search by multi-scale spectral residue analysis

Using an object detector over a whole image can require significant processing time. This is so since the majority of the images, in common scenarios, is composed of non-trivial amounts of background information, such as sky, ground and water. To alleviate this computational load, image search space reduction methods can make the detection procedure focus on more distinctive image regions. In this sense, we propose here the use of saliency information to organize regions based on their probability of containing objects. The proposed method was grounded on a multi-scale spectral residue (MSR) analysis for saliency detection. For better search space reduction, our method enables fine control of search scale, presents more robustness to variations on saliency intensity along an object length, and also a straightforward way to control the balance between search space reduction and false negatives, both being a consequence of region selection. MSR was capable of making object detection three to five times faster compared to the same detector without MSR. A thorough analysis was accomplished to demonstrate the effectiveness of the proposed method using a custom LabelMe dataset of person images, and also a Pascal VOC 2007 dataset, containing several distinct object classes.

Evaluation of the Huffman Encoding for Memory Optimization on Hardware Network Intrusion Detection

The design of specialized hardware for Network Intrusion Detection has been subject of intense research over the last decade due to its considerably higher performance compared to software implementations. In this context, one of the limiting factors is the finite amount of memory resources versus the increasing number of threat patterns to be analyzed. This paper proposes an architecture based on the Huffman algorithm for encoding, storage and decoding of these patterns in order to optimize such resources. We have made tests with simulation and synthesis in FPGA of rule subsets of the Snort software, and analysis indicate a saving of up to 73 percent of the embedded memory resources of the chip.