Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00058
L. Pacífico, Teresa B Ludermir, Larissa F. S. Britto
Image segmentation is a fundamental process for image analysis and computer vision. One of the most popular image segmentation methods is Otsu algorithm, originally proposed to segment a grayscale image in two classes, but extended to multi-level thresholding afterwards. Although effective, the computational cost for multi-level Otsu limits its application in real world problems, and, recently, many evolutionary algorithms (EAs) have been applied to optimize Otsu algorithm. In this paper, a hybrid Otsu and improved Group Search Optimization (GSO) algorithm is presented to deal with multi-level color image thresholding problem, named IGSO. IGSO implements a weedout operator to prune the worst members from GSO population. We also evaluate the influence of two treatments to deal with outbounded individuals from EAs population. The proposed IGSO is compared to other EAs from literature through twelve real color image problems, showing its potential and robustness even when compared to original GSO algorithm.
{"title":"A Hybrid Improved Group Search Optimization and Otsu Method for Color Image Segmentation","authors":"L. Pacífico, Teresa B Ludermir, Larissa F. S. Britto","doi":"10.1109/BRACIS.2018.00058","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00058","url":null,"abstract":"Image segmentation is a fundamental process for image analysis and computer vision. One of the most popular image segmentation methods is Otsu algorithm, originally proposed to segment a grayscale image in two classes, but extended to multi-level thresholding afterwards. Although effective, the computational cost for multi-level Otsu limits its application in real world problems, and, recently, many evolutionary algorithms (EAs) have been applied to optimize Otsu algorithm. In this paper, a hybrid Otsu and improved Group Search Optimization (GSO) algorithm is presented to deal with multi-level color image thresholding problem, named IGSO. IGSO implements a weedout operator to prune the worst members from GSO population. We also evaluate the influence of two treatments to deal with outbounded individuals from EAs population. The proposed IGSO is compared to other EAs from literature through twelve real color image problems, showing its potential and robustness even when compared to original GSO algorithm.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129900714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00103
Diego Minatel, Alan Valejo, A. Lopes
The popularization of GPS has generated a massive amount of geographic data organized in trajectories. Trajectories are ordered sequences of geographic points that represent a path of any moving object, which provides information on the mobility behavior of this moving objects. To improve the understanding of trajectories, places of greater importance are referred to as stay points and indicate that a user has remained in this correspondent place for a significant time. In the literature, stay points are commonly represented through networks to facilitate trajectory mining. Nonetheless, to the best of our knowledge, there is a lack of studies addressing the quality of these networks. This article addresses this gap and presents a network construction analysis through stay points by using external validity criteria.
{"title":"Trajectory Network Assessment Based on Analysis of Stay Points Cluster","authors":"Diego Minatel, Alan Valejo, A. Lopes","doi":"10.1109/BRACIS.2018.00103","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00103","url":null,"abstract":"The popularization of GPS has generated a massive amount of geographic data organized in trajectories. Trajectories are ordered sequences of geographic points that represent a path of any moving object, which provides information on the mobility behavior of this moving objects. To improve the understanding of trajectories, places of greater importance are referred to as stay points and indicate that a user has remained in this correspondent place for a significant time. In the literature, stay points are commonly represented through networks to facilitate trajectory mining. Nonetheless, to the best of our knowledge, there is a lack of studies addressing the quality of these networks. This article addresses this gap and presents a network construction analysis through stay points by using external validity criteria.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121305487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00043
M. L. D. Dias, A. Freire, A. H. S. Júnior, A. Neto, J. Gomes
The Minimal Learning Machine (MLM) is a supervised method in which learning consists of fitting a multiresponse linear regression model between distances computed from the input and output spaces. A critical issue related to the training process in MLMs is the selection of prototypes, also called reference points (RPs), from which distances are taken. In its original formulation, the MLM selects the RPs randomly from the data. In this paper we empirically show that the original random selection may lead to a poor generalization capability. In addition, we propose a novel pruning method for selecting RPs based on L_1/2 norm regularization. Our results show that the proposed method is able to outperform the original MLM and its variants.
{"title":"Sparse Minimal Learning Machines Via L_1/2 Norm Regularization","authors":"M. L. D. Dias, A. Freire, A. H. S. Júnior, A. Neto, J. Gomes","doi":"10.1109/BRACIS.2018.00043","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00043","url":null,"abstract":"The Minimal Learning Machine (MLM) is a supervised method in which learning consists of fitting a multiresponse linear regression model between distances computed from the input and output spaces. A critical issue related to the training process in MLMs is the selection of prototypes, also called reference points (RPs), from which distances are taken. In its original formulation, the MLM selects the RPs randomly from the data. In this paper we empirically show that the original random selection may lead to a poor generalization capability. In addition, we propose a novel pruning method for selecting RPs based on L_1/2 norm regularization. Our results show that the proposed method is able to outperform the original MLM and its variants.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121493427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/bracis.2018.00096
Milton Condori Fernández, Leliane Nunes de Barros, Karina Valdivia Delgado
The most efficient approach to solve probabilistic planning problems is based on stochastic shortest path (SSP) and uses heuristic search to find a policy that minimizes the expected accumulated cost to the goal (MINCOST criterion). However, this approach can only solve problems with dead ends (states from which it is not possible to reach the goal) if an efficient dead end detection heuristic is used. Another solution would be to plan in two phases: maximizing the probability to reach the goal (MAXPROB) and then minimizing the expected cost (MINCOST). While there exist several heuristics to solve MINCOST, there is not known efficient heuristics to solve MAXPROB. A recent work proposes the first heuristic that takes into account the probabilities, called h pom, which solves a relaxed version of an SSP as a linear program in the dual space. However, to solve large problems with dead ends, h pom must be augmented with a dead end detection heuristic (e.g., h_pom and h_max ). In this work, we propose two new heuristics based on h pom. The first, h^pe_pom (s), estimates the minimal cost of state s to reach the goal, including new variables and constraints for the dead ends and adding an expected penalty for reaching them. The second, h ppom (s), estimates the maximum probability of state s to reach the goal, and is used to solve MAXPROB problems by ignoring action costs; We claim that h ppom (s) is the first heuristic for MAXPROB. Empirical results show that h^pe_pom can solve larger planning instances when compared to h pom h_max.
{"title":"Occupation Measure Heuristics to Solve Stochastic Shortest Path with Dead Ends","authors":"Milton Condori Fernández, Leliane Nunes de Barros, Karina Valdivia Delgado","doi":"10.1109/bracis.2018.00096","DOIUrl":"https://doi.org/10.1109/bracis.2018.00096","url":null,"abstract":"The most efficient approach to solve probabilistic planning problems is based on stochastic shortest path (SSP) and uses heuristic search to find a policy that minimizes the expected accumulated cost to the goal (MINCOST criterion). However, this approach can only solve problems with dead ends (states from which it is not possible to reach the goal) if an efficient dead end detection heuristic is used. Another solution would be to plan in two phases: maximizing the probability to reach the goal (MAXPROB) and then minimizing the expected cost (MINCOST). While there exist several heuristics to solve MINCOST, there is not known efficient heuristics to solve MAXPROB. A recent work proposes the first heuristic that takes into account the probabilities, called h pom, which solves a relaxed version of an SSP as a linear program in the dual space. However, to solve large problems with dead ends, h pom must be augmented with a dead end detection heuristic (e.g., h_pom and h_max ). In this work, we propose two new heuristics based on h pom. The first, h^pe_pom (s), estimates the minimal cost of state s to reach the goal, including new variables and constraints for the dead ends and adding an expected penalty for reaching them. The second, h ppom (s), estimates the maximum probability of state s to reach the goal, and is used to solve MAXPROB problems by ignoring action costs; We claim that h ppom (s) is the first heuristic for MAXPROB. Empirical results show that h^pe_pom can solve larger planning instances when compared to h pom h_max.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"133 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127313254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00059
V. P. L. Varela, Arthur Oliveira, Paulo Rodrigues, Miller Horvath
The Firefly Algorithm (FA) is a meta-heuristic optimization algorithm that mimics the social behaviour of fireflies. The FA is suggested as a good algorithm for tracking, but it still requires too much computational process. This study propose a different approach using the FA as a Tracking Algorithm by using Tsallis Entropy and qFA thresholds from the previous frame as heuristic for the next frame to enhance its speed.
{"title":"qFA: An Optimized Based-Tracking Approach Using Firefly Algorithm","authors":"V. P. L. Varela, Arthur Oliveira, Paulo Rodrigues, Miller Horvath","doi":"10.1109/BRACIS.2018.00059","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00059","url":null,"abstract":"The Firefly Algorithm (FA) is a meta-heuristic optimization algorithm that mimics the social behaviour of fireflies. The FA is suggested as a good algorithm for tracking, but it still requires too much computational process. This study propose a different approach using the FA as a Tracking Algorithm by using Tsallis Entropy and qFA thresholds from the previous frame as heuristic for the next frame to enhance its speed.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130871052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00080
G. Oliveira, R. G. O. Silva, Laurence Rodrigues do Amaral, L. G. A. Martins
Formation control is the task of coordinating a group of robots to get into and to maintain a formation with a specific shape while moving in the environment. In this work we investigated models based on cellular automata applied to this task. We implemented methods previously described in the literature and found some limitations. Thus, an evolutionary-cooperative method is proposed, using the search power of a genetic algorithm along with the compact rules and simplified processing of cellular automata. The proposal required low computational infrastructure and was tested in a robotics simulator (Webots) with a team of 3 e-puck robots. The new model exhibited a better behaviour than their precursors in several scenarios, improving the robot's trajectory and formation maintenance.
{"title":"An Evolutionary-Cooperative Model Based on Cellular Automata and Genetic Algorithms for the Navigation of Robots Under Formation Control","authors":"G. Oliveira, R. G. O. Silva, Laurence Rodrigues do Amaral, L. G. A. Martins","doi":"10.1109/BRACIS.2018.00080","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00080","url":null,"abstract":"Formation control is the task of coordinating a group of robots to get into and to maintain a formation with a specific shape while moving in the environment. In this work we investigated models based on cellular automata applied to this task. We implemented methods previously described in the literature and found some limitations. Thus, an evolutionary-cooperative method is proposed, using the search power of a genetic algorithm along with the compact rules and simplified processing of cellular automata. The proposal required low computational infrastructure and was tested in a robotics simulator (Webots) with a team of 3 e-puck robots. The new model exhibited a better behaviour than their precursors in several scenarios, improving the robot's trajectory and formation maintenance.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132843040","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00063
Samuel E. L. Oliveira, P. Brum, A. Lacerda, G. Pappa
Group recommendation methods deal with scenarios where a group is the target of recommendation instead of a single user. An initial approach followed by these methods was to aggregate the rankings generated to each individual user of the group by traditional recommender systems. This approach was replaced to more sophisticated methods, but the potential and simplicity of the aggregation strategies were underexplored. This paper proposes to use multiple recommenders to generate recommendations to single group members before aggregating their recommendations. We show that this strategy significantly improves the results of aggregating single recommenders while overcoming the problem of selecting the best recommendation algorithm. We also propose five heuristics to select a subset of the available recommenders to be aggregated. We tested heuristics in seven dataset variations, showing that by using half of the available algorithms we can achieve results similar or better than those obtained by the whole set.
{"title":"Exploiting Multiple Recommenders to Improve Group Recommendation","authors":"Samuel E. L. Oliveira, P. Brum, A. Lacerda, G. Pappa","doi":"10.1109/BRACIS.2018.00063","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00063","url":null,"abstract":"Group recommendation methods deal with scenarios where a group is the target of recommendation instead of a single user. An initial approach followed by these methods was to aggregate the rankings generated to each individual user of the group by traditional recommender systems. This approach was replaced to more sophisticated methods, but the potential and simplicity of the aggregation strategies were underexplored. This paper proposes to use multiple recommenders to generate recommendations to single group members before aggregating their recommendations. We show that this strategy significantly improves the results of aggregating single recommenders while overcoming the problem of selecting the best recommendation algorithm. We also propose five heuristics to select a subset of the available recommenders to be aggregated. We tested heuristics in seven dataset variations, showing that by using half of the available algorithms we can achieve results similar or better than those obtained by the whole set.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124161654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00077
Vinicius M. A. Souza, T. P. D. Silva, Gustavo E. A. P. A. Batista
In general, data stream classifiers consider that the actual label of every unlabeled instance is available immediately after it issues a classification. The immediate availability of class labels allows the supervised monitoring of the data distribution and the error rate to verify whether the current classifier is outdated. Further, if a change is detected, the classifier has access to all recent labeled data to update the model. However, this assumption is very optimistic for most (if not all) applications. Given the costs and labor involved to obtain labels, failures in data acquisition or restrictions of the classification problem, a more reasonable assumption would be to consider the delayed availability of class labels. In this paper, we experimentally analyze the impact of latency on the performance of stream classifiers and call the attention of the community for the need to consider this critical variable in the evaluation process. We also make suggestions to avoid possible biased conclusions due to ignoring the delayed nature of stream problems. These are relevant contributions since few studies consider this variable in new algorithms proposals. Also, we propose a new evaluation measure (Kappa-Latency) that takes into account the arrival delay of actual labels to evaluate and compare a set of classifiers.
{"title":"Evaluating Stream Classifiers with Delayed Labels Information","authors":"Vinicius M. A. Souza, T. P. D. Silva, Gustavo E. A. P. A. Batista","doi":"10.1109/BRACIS.2018.00077","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00077","url":null,"abstract":"In general, data stream classifiers consider that the actual label of every unlabeled instance is available immediately after it issues a classification. The immediate availability of class labels allows the supervised monitoring of the data distribution and the error rate to verify whether the current classifier is outdated. Further, if a change is detected, the classifier has access to all recent labeled data to update the model. However, this assumption is very optimistic for most (if not all) applications. Given the costs and labor involved to obtain labels, failures in data acquisition or restrictions of the classification problem, a more reasonable assumption would be to consider the delayed availability of class labels. In this paper, we experimentally analyze the impact of latency on the performance of stream classifiers and call the attention of the community for the need to consider this critical variable in the evaluation process. We also make suggestions to avoid possible biased conclusions due to ignoring the delayed nature of stream problems. These are relevant contributions since few studies consider this variable in new algorithms proposals. Also, we propose a new evaluation measure (Kappa-Latency) that takes into account the arrival delay of actual labels to evaluate and compare a set of classifiers.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116629437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00049
Églen Protas, José Douglas Bratti, P. Ribeiro, Paulo L. J. Drews-Jr, S. Botelho
Convolutional Neural Networks became a state-of-the-art approach for many different problems of computer vision, pattern recognition, and image processing. However, due to the large number of parameters of these architectures, researchers may find difficult to explain what the networks are using as discriminative patterns. An alternative to better understand the behavior of the learned convolutional kernels is the use of visualization techniques. Currently, visualization techniques are more frequently applied to classification tasks. In this paper, we address the visualization of image-to-image translation. One of the contributions of our work is the possibility to modify a network based on the kernel visualization and achieve superior results.
{"title":"Visualization Techniques Applied to Image-to-Image Translation","authors":"Églen Protas, José Douglas Bratti, P. Ribeiro, Paulo L. J. Drews-Jr, S. Botelho","doi":"10.1109/BRACIS.2018.00049","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00049","url":null,"abstract":"Convolutional Neural Networks became a state-of-the-art approach for many different problems of computer vision, pattern recognition, and image processing. However, due to the large number of parameters of these architectures, researchers may find difficult to explain what the networks are using as discriminative patterns. An alternative to better understand the behavior of the learned convolutional kernels is the use of visualization techniques. Currently, visualization techniques are more frequently applied to classification tasks. In this paper, we address the visualization of image-to-image translation. One of the contributions of our work is the possibility to modify a network based on the kernel visualization and achieve superior results.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116521378","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-10-01DOI: 10.1109/BRACIS.2018.00021
R. Berri, F. Osório
In this work, a nonintrusive system has been developed using features from inertial sensors, car telemetry, and road lane data, enabling to recognize the driving style of a drunk driver. Drunk drivers caused 10,497 deaths on USA roads in 2016 according to NHTSA. The Naturalistic Driver Behavior Dataset (NDBD) was created specifically for this work and it was used to test the proposed system. The proposed system was designed to study drunk driving situations, but it can also be used to detect any other psychoactive drugs consumption that causes abnormal driver behaviors during driving. The classifier system's output is "no risk" (normal driving) or "risk" (drunk/abnormal driving). If the system is connected to an autonomous or semi-autonomous car control system, it can be enabled to step in and act in order to avoid dangerous situations, or it can activate an alarm, or also ask for external help (e.g. contact authorities). The best results achieved in the experiments obtained 98% of accuracy in NDBD frames and only 1.5% of frames labeled in NDBD as "no risk" had a wrong prediction. The proposed system is composed by an MLP neural classifier using sigmoidal activation function and with 14 neurons in input layer, 18 neurons in hidden layer, and 1 neuron in output layer of the network. It uses periods of 220 frames (22 seconds) for the predictions and a buffer of the last 3 predictions was used for reducing the number of false predictions for "risk" output. Thus, it could avoid wrong predictions (false positives), avoiding to incorrectly enable the alarms and semi-autonomous car control system.
{"title":"A Nonintrusive System for Detecting Drunk Drivers in Modern Vehicles","authors":"R. Berri, F. Osório","doi":"10.1109/BRACIS.2018.00021","DOIUrl":"https://doi.org/10.1109/BRACIS.2018.00021","url":null,"abstract":"In this work, a nonintrusive system has been developed using features from inertial sensors, car telemetry, and road lane data, enabling to recognize the driving style of a drunk driver. Drunk drivers caused 10,497 deaths on USA roads in 2016 according to NHTSA. The Naturalistic Driver Behavior Dataset (NDBD) was created specifically for this work and it was used to test the proposed system. The proposed system was designed to study drunk driving situations, but it can also be used to detect any other psychoactive drugs consumption that causes abnormal driver behaviors during driving. The classifier system's output is \"no risk\" (normal driving) or \"risk\" (drunk/abnormal driving). If the system is connected to an autonomous or semi-autonomous car control system, it can be enabled to step in and act in order to avoid dangerous situations, or it can activate an alarm, or also ask for external help (e.g. contact authorities). The best results achieved in the experiments obtained 98% of accuracy in NDBD frames and only 1.5% of frames labeled in NDBD as \"no risk\" had a wrong prediction. The proposed system is composed by an MLP neural classifier using sigmoidal activation function and with 14 neurons in input layer, 18 neurons in hidden layer, and 1 neuron in output layer of the network. It uses periods of 220 frames (22 seconds) for the predictions and a buffer of the last 3 predictions was used for reducing the number of false predictions for \"risk\" output. Thus, it could avoid wrong predictions (false positives), avoiding to incorrectly enable the alarms and semi-autonomous car control system.","PeriodicalId":405190,"journal":{"name":"2018 7th Brazilian Conference on Intelligent Systems (BRACIS)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125095799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：