Pedestrian Detection and Tracking Using Deformable Part Models and Kalman Filtering

Abstract： Pedestrian detection techniques are important and challenging especially for complex real world scenes. They can be used for ensuring pedestrian safety, ADASs （advance driver assistance systems） and safety surveillance systems. In this paper, we propose a novel approach for multi-person tracking-by-detection using deformable part models in Kalman filtering framework. The Kalman filter is used to keep track of each person and a unique label is assigned to each tracked individual. Based on this approach, people can enter and leave the scene randomly. We test and demonstrate our results on Caltech Pedestrian benchmark, which is two orders of magnitude larger than any other existing datasets and consists of pedestrians varying widely in appearance, pose and scale. Complex situations such as people occluded by each other are handled gracefully and individual persons can be tracked correctly after a group of people split. Experiments confirm the real-time performance and robustness of our system, working in complex scenes. Our tracking model gives a tracking accuracy of 72.8% and a tracking precision of 82.3%. We can further reduce false positives by 2.8%, using Kalman filtering.     

Automated analysis of images in documents for intelligent document search

Authors use images to present a wide variety of important information in documents. For example, two-dimensional (2-D) plots display important data in scientific publications. Often, end-users seek to extract this data and convert it into a machine-processible form so that the data can be analyzed automatically or compared with other existing data. Existing document data extraction tools are semi-automatic and require users to provide metadata and interactively extract the data. In this paper, we describe a system that extracts data from documents fully automatically, completely eliminating the need for human intervention. The system uses a supervised learning-based algorithm to classify figures in digital documents into five classes: photographs, 2-D plots, 3-D plots, diagrams, and others. Then, an integrated algorithm is used to extract numerical data from data points and lines in the 2-D plot images along with the axes and their labels, the data symbols in the figure’s legend and their associated labels. We demonstrate that the proposed system and its component algorithms are effective via an empirical evaluation. Our data extraction system has the potential to be a vital component in high volume digital libraries.     

Supply chain risk mitigation competency: an individual-level knowledge-based perspective

Supply chain managers are responsible for making decisions regarding supply chain risk in order to mitigate the impact of supply chain disruptions. This study develops and tests a theoretical model that leverages the individual-level knowledge-based view perspective to understand the process through which risk mitigation orientation of the supply chain manager contributes to his/her absorptive capacity. A supply chain manager’s absorptive capacity, in turn, enhances his/her ability to effectively mitigate supply chain risk. Study findings demonstrate that supply chain managers with high-risk mitigation orientation have greater level of absorptive capacity which enhances their risk mitigation competency. This study represents the first development and testing of a model that examines individual-level knowledge management factors that affect supply chain risk mitigation competency. This research emphasises the importance of the individual supply chain manager in managing risk and illustrates how theoretical perspectives from the knowledge management, supply chain risk and organisational behaviour literature can be fruitfully adopted to explain behaviour in the field of supply chain risk management.     

Intelligent System Application to Monitor the Smart City Building Lighting

A smart city incorporates infrastructure methods that are environmentally responsible, such as smart communications, smart grids, smart energy, and smart buildings. The city administration has prioritized the use of cutting-edge technology and informatics as the primary strategy for enhancing service quality, with energy resources taking precedence. To achieve optimal energy management in the multidimensional system of a city tribe, it is necessary not only to identify and study the vast majority of energy elements, but also to define their implicit interdependencies. This is because optimal energy management is required to reach this objective. The lighting index is an essential consideration when evaluating the comfort indicators. In order to realize the concept of a smart city, the primary objective of this research is to create a system for managing and monitoring the lighting index. It is possible to identify two distinct phases within the intelligent system. Once data collection concludes, the monitoring system will be activated. In the second step, the operation of the control system is analyzed and its effect on the performance of the numerical model is determined. This evaluation is based on the proposed methodology. The optimized results were deemed satisfactory because they maintained the brightness index value (79%) while consuming less energy. The intelligent implementation system generated satisfactory outcomes, which were observed 1.75 times on average.     

An NTU- model for alternative refrigerants

This paper presents a steady state simulation model to predict the performance of alternative refrigerants in vapour compression refrigeration/heat pump systems. The model is based on the NTU- method in analysing the heat exchangers following an elemental approach. The model extends its applicability to new refrigerants including hydrocarbons and uses a large database of REFPROP package for refrigerant properties. The main inputs to the model include the physical details of the heat exchangers, compressor efficiency, mass flow rates of heat transfer fluids and their inlet temperatures to the evaporator and the condenser, the pressure drops across the heat exchangers and the capacity of either the evaporator or condenser (in kW). The model results are validated with a wide range of experimental data of HCFC22 and propane (HC290) on a heat pump test facility for a number of parameters, e.g. coefficient of performance, condenser capacity, mass flow rate of the refrigerant and compressor discharge temperature. Although the model is currently tested for pure refrigerants using compact brazed plate (counter flow type) heat exchangers, it can also be applied to mixture of refrigerants as well as to other types of heat exchangers.

Résumé

Dans cet article, on présente un modèle de simulation de régime permanent pour prédire la performance des frigorigènes de remplacement dans les systèmes frigorifiques ou les pompes à chaleur à compression de vapeur. Fondé sur la méthode NTU- utilisée pour analyser les échangeurs de chaleur, ce modèle emploie une approche élémentaire. Ce modèle étend la méthode aux nouveaux frigorigènes, y compris deees hydrocarbures, et utilise une base de données étendue, celle de REFPROP, pour les propriétés des frigorigènes. Les principaux paramètres du modèle comprennent des détails physiques sur les échangeurs de chaleur, le rendement des compresseurs, et les débits massiques des fluides de transfert de chaleur et leurs températures à l'entrée de l'évaporateur ou du condenseur, la chute de pression à travers les échangeurs de chaleur et la puissance soit de l'évaporateur, soit du condenseur (exprimés en kW). Les résultats obtensus en utilisant ce modèle sont validés pour une large gamme de données expérimentales obtenus avec le HCFC22 et avec le propane (le HC290) sur un banc d'essai de pompe à chaleur et pour un certain nombre de paramètres, par exemple le coefficient de performance, la puissance du compresseur, le débit massique du frigorigène et la température du frigorigène à la sortie du compresseur. En ce moment, le comportement des frigorigènes purs utilisés dans des échangeurs de chaleur compacts à plaques brasées (de type contre-courant) est en train d'être étudié; le modèle peut également être appliqué aux mélanges de frigorigènes et à d'autres types d'échangeurs de chaleur.     

Artificial intelligence applications in Permanent Magnet Brushless DC motor drives

Permanent Magnet Brushless DC (PMBLDC) machines are more popular due its simple structure and low cost. Improvements in permanent magnetic materials and power electronic devices have resulted in reliable, cost effective PMBLDC drives, for many applications. Advances in artificial intelligent applications like neural network, fuzzy logic, Genetic algorithm etc. have made tremendous impact on electric motor drives. The brushless DC motor is a multivariable and non-linear system. In conventional PMBLDC drives speed and position sensing of brushless DC motors require high degree of accuracy. Unfortunately, traditional methods of control require detailed modelling of all the motor parameters to achieve this. The Intelligent control techniques like, fuzzy logic control/Neural network control etc. uses heuristic input–output relations to deal with vague and complex situations. This paper presents a literature survey on the intelligent control techniques for PMBLDC motor drives. Various AI techniques for PMBLDC motor drives are described. Attempt is made to provide a guideline and quick reference for the researchers and practicing engineers those are working in the area of PMBLDC motor drives.     

Privacy preserving Back-propagation neural network learning over arbitrarily partitioned data

Neural networks have been an active research area for decades. However, privacy bothers many when the training dataset for the neural networks is distributed between two parties, which is quite common nowadays. Existing cryptographic approaches such as secure scalar product protocol provide a secure way for neural network learning when the training dataset is vertically partitioned. In this paper, we present a privacy preserving algorithm for the neural network learning when the dataset is arbitrarily partitioned between the two parties. We show that our algorithm is very secure and leaks no knowledge (except the final weights learned by both parties) about other party’s data. We demonstrate the efficiency of our algorithm by experiments on real world data.     

Speed scaling with a solar cell

We consider the setting of a device that obtains its energy from a battery and some regenerative source such as a solar cell. We consider the speed scaling problem of scheduling a collection of tasks with release times, deadlines, and sizes, so as to minimize the energy recharge rate of the regenerative source. This is the first theoretical investigation of speed scaling for devices with a regenerative energy source. We show that the problem can be expressed as a polynomial sized convex program. We show that, using the KKT conditions, one can obtain an efficient algorithm to verify the optimality of a schedule. We show that the energy optimal YDS schedule is 2-approximate with respect to the recharge rate. We show that the online algorithm BKP is O(1)$O\left(1\right)$-competitive with respect to recharge rate.