Publications

Here, my peer-reviewed journal articles and conference papers are listed chronologically. For details you may check my Google Scholar page.

Journal Articles

Active eye-in-hand data management to improve the robotic object detection performance

Journal: Computers

Authors: P. Hoseini, J. Blankenburg, M. Nicolescu, M. Nicolescu, D. Feil-Seifer

Year: 2019

Link: MDPI or pdf

Note: Article appeared on the issue cover of December 2019 of the quarterly journal (The featured image)

Abstract:

Adding to the number of sources of sensory information can be efficacious in enhancing the object detection capability of robots. In the realm of vision-based object detection, in addition to improving the general detection performance, observing objects of interest from different points of view can be central to handling occlusions. In this paper, a robotic vision system is proposed that constantly uses a 3D camera, while actively switching to make use of a second RGB camera in cases where it is necessary. The proposed system detects objects in the view seen by the 3D camera, which is mounted on a humanoid robot’s head, and computes a confidence measure for its recognitions. In the event of low confidence regarding the correctness of the detection, the secondary camera, which is installed on the robot’s arm, is moved toward the object to obtain another perspective of the object. With the objects detected in the scene viewed by the hand camera, they are matched to the detections of the head camera, and subsequently, their recognition decisions are fused together. The decision fusion method is a novel approach based on the Dempster–Shafer evidence theory. Significant improvements in object detection performance are observed after employing the proposed active vision system.

High-speed general purpose genetic algorithm processor

Journal: IEEE Transactions on Cybernetics

Authors: P. Hoseini, S. Moshfe, M. Saber Zaeimian, A. Khoei, K. Hadidi

Year: 2016

Link: IEEE or pdf

Abstract:

In this paper, an ultrafast steady-state genetic algorithm processor (GAP) is presented. Due to the heavy computational load of genetic algorithms (GAs), they usually take a long time to find optimum solutions. Hardware implementation is a significant approach to overcome the problem by speeding up the GAs procedure. Hence, we designed a digital CMOS implementation of GA in 0.18 μm process. The proposed processor is not bounded to a specific application. Indeed, it is a general-purpose processor, which is capable of performing optimization in any possible application. Utilizing speed-boosting techniques, such as pipeline scheme, parallel coarse-grained processing, parallel fitness computation, parallel selection of parents, dual-population scheme, and support for pipelined fitness computation, the proposed processor significantly reduces the processing time. Furthermore, by relying on a built-in discard operator the proposed hardware may be used in constrained problems that are very common in control applications. In the proposed design, a large search space is achievable through the bit string length extension of individuals in the genetic population by connecting the 32-bit GAPs. In addition, the proposed processor supports parallel processing, in which the GAs procedure can be run on several connected processors simultaneously.

A fully programmable analog CMOS rational-powered membership function generator with continuously adjustable high precision parameters

Journal: Circuits, Systems, and Signal Processing

Authors: S. Moshfe, P. Hoseini, A. Khoei, K. Hadidi

Year: 2014

Link: Springer or pdf

Abstract:

This paper presents a novel structure for implementing rational-powered membership functions (RPMFs), which are the extended forms of triangular/trapezoidal membership functions and those functions which are generated by applying linguistic hedges. The hardware realization of an RPMF consists of a triangular membership function generator circuit followed by a rational-powered generator module (RPGM). A novel fully programmable compact triangular/trapezoidal/s-shaped/z-shaped membership function generator with the ability to continuously change parameters is presented which is compatible with the proposed RPGM. A new method is introduced to implement the RPGM based on the approximation of the function “x a” by the functions square and square-rooter which are simply implemented in a current-mode analog approach based on the translinear principle, which leads to a design that is simple, and has high accuracy and less hardware usage, with a resulting lower chip area and lower power consumption. The designed circuit was simulated by an HSPICE simulator with level 49 parameters (BSIM3v3), and the simulation results show that the maximum power consumption of the RPGM is 800 μW, while the maximum RMS error is 1.25 %. Finally, layouts of the circuits prepared using Cadence software are presented.

Efficient contrast enhancement of images using hybrid ant colony optimisation, genetic algorithm, and simulated annealing

Journal: Digital Signal Processing

Authors: P. Hoseini, M. G. Shayesteh

Year: 2013

Link: Elsevier or pdf

Abstract:

In this paper, we propose a hybrid algorithm including Genetic Algorithm (GA), Ant Colony Optimisation (ACO), and Simulated Annealing (SA) metaheuristics for increasing the contrast of images. In this way, contrast enhancement is obtained by global transformation of the input intensities. Ant colony optimisation is used to generate the transfer functions which map the input intensities to the output intensities. Simulated annealing as a local search method is utilised to modify the transfer functions generated by ant colony optimisation. And genetic algorithm has the responsibility of evolutionary process of antsʼ characteristics. The employed fitness function operates automatically and tends to provide a balance between contrast and naturalness of images. The results indicate that the new method achieves images with higher contrast than the previously presented methods from the subjective and objective viewpoints. Further, the proposed algorithm preserves the natural look of input images.

Conference Papers

Collaborative human-robot hierarchical task execution with an activation spreading architecture

Conference: International Conference on Social Robotics

Authors: B. A. Anima, J. Blankenburg, M. Zagainova, P. Hoseini, M. Chowdhury, D. Feil-Seifer, M. Nicolescu, M. Nicolescu

Year: 2019

Link: Springer or pdf

Note: Best paper award finalist

Abstract:

This paper addresses the problem of human-robot task execution for hierarchical task plans. The main contributions are the ability for dynamic allocation of tasks in human-robot teams and opportunistic task execution given different environmental conditions. The human-robot collaborative task is represented in a tree structure which consists of sequential, non-ordering, and alternative paths of execution. The general approach to enable human-robot collaborative task execution is to have the robot maintain an updated, simulated version of the human’s task representation, which is similar to the robot’s own controller for the same task. Continuous peer node message passing between the agents’ task representations enables both to coordinate their task execution, so that they perform the task given its required execution constraints and they do not both work on the same task component. A tea-table task scenario was designed for validation with overlapping and non-overlapping sub-tasks between a human and a Baxter robot.

An active robotic vision system with a pair of moving and stationary cameras

Conference: International Symposium on Visual Computing

Authors: P. Hoseini, J. Blankenburg, M. Nicolescu, M. Nicolescu, D. Feil-Seifer

Year: 2019

Link: Springer or pdf

Abstract:

Vision is one of the main potential sources of information for robots to understand their surroundings. For a vision system, a clear and close enough view of objects or events, as well as the viewpoint angle can be decisive in obtaining useful features for the vision task. In order to prevent performance drops caused by inefficient camera orientations and positions, manipulating cameras, which falls under the domain of active perception, can be a viable option in a robotic environment.

In this paper, a robotic object detection system is proposed that is capable of determining the confidence of recognition after detecting objects in a camera view. In the event of a low confidence, a secondary camera is moved toward the object and performs an independent detection round. After matching the objects in the two camera views and fusing their classification decisions through a novel transferable belief model, the final detection results are obtained. Real world experiments show the efficacy of the proposed approach in improving the object detection performance, especially in the presence of occlusion.

Active object detection through dynamic incorporation of Dempster-Shafer fusion for robotic applications

Conference: International Conference on Vision, Image and Signal Processing

Authors: P. Hoseini, M. Nicolescu, M. Nicolescu

Year: 2018

Link: ACM or pdf

Abstract:

Employing multiple sensing capabilities in a robotic platform offers significant advantages in increasing the recognition abilities of robots. Specifically, for vision-based object detection in a real-world environment, acquiring information from different viewpoints might be decisive for correct classifications in the presence of occlusions or to disambiguate between similar objects. For this reason, an active vision object detection system is proposed in this paper. It is implemented on a robotic environment that uses a 3D camera mounted on the robot head and an RGB camera on its hand. The system tries to detect and recognize objects being seen from the head camera, while computing a confidence score on the classification. In the case of an unreliable classification, another stage of object recognition is dynamically requested, but this time from the viewpoint of the hand camera. The objects detected from the two cameras are matched and their classification decisions are fused through a novel fusion approach based on the Dempster-Shafer evidence theory. Experimental results show sizable improvements in object recognition performance compared to a traditional singlecamera configuration, as well as applicability to handling partial occlusions.

Handling ambiguous object recognition situations in a robotic environment via dynamic information fusion

Conference: IEEE Conference on Cognitive and Computational Aspects of Situation Management

Authors: P. Hoseini, M. Nicolescu, M. Nicolescu

Year: 2018

Link: IEEE or pdf

Note: Won the student travel grant

Abstract:

Vision is usually a rich source of information for robots aiming to understand activities that take place in their surroundings, where a relevant task can be to detect and recognize objects of interest. In real world conditions a robot may not have a good viewing angle or be sufficiently close to an object to distinguish its features, which can lead to misclassifications. One solution to address this problem is active vision, leading to an improved level of situational awareness in a dynamic environment. In that context, a vision system on the robot actively manipulates the camera to obtain enough discriminating features for the task of object detection and recognition. In this paper, an active vision system is proposed that is able to identify a situation with a high possibility of misclassification (for example, partial occlusions) and then to take appropriate action by dynamically incorporating another camera installed on the robot's hand. A decision fusion technique based on a transferable belief model generates the final classification results. Experimental results show considerable improvements in object detection and recognition performance.

A distributed control architecture for collaborative multi-robot task allocation

Conference: International Conference on Humanoid Robotics

Authors: J. Blankenburg, S. B. Banisetty, P. Hoseini, L. Fraser, D. Feil-Seifer, M. Nicolescu, M. Nicolescu

Year: 2017

Link: IEEE or pdf

Abstract:

This paper addresses the problem of task allocation for multi-robot systems that perform tasks with complex, hierarchical representations which contain different types of ordering constraints and multiple paths of execution. We propose a distributed multi-robot control architecture that addresses the above challenges and makes the following contributions: i) it allows for on-line, dynamic allocation of robots to various steps of the task, ii) it ensures that the collaborative robot system will obey all of the task constraints and iii) it allows for opportunistic, flexible task execution given different environmental conditions. This architecture uses a distributed messaging system to allow the robots to communicate. Each robot uses its own state and team member states to keep track of the progress on a given task and identify which subtasks to perform next using an activation spreading mechanism. We demonstrate the proposed architecture on a team of two humanoid robots (a PR2 and a Baxter) performing hierarchical tasks.

A modified steady state genetic algorithm suitable for fast pipelined hardware

Conference: IEEE Congress on Evolutionary Computation

Authors: P. Hoseini, S. Moshfe, S. J. Louis, M. Nicolescu

Year: 2017

Link: IEEE or pdf

Abstract:

In this paper, a modification of steady state genetic algorithm, called dual-population scheme, is proposed to improve its execution speed on electronic hardware. It utilizes two memories to store two interactive populations on them. The system, inherently interchanges chromosomes between these populations. In this manner, it can fully benefit from the pipeline processing on hardware. It is shown that the proposed method performs much faster than the standard steady state and canonical genetic algorithms on pipelined genetic hardware. Moreover, the searching performance, repeatability and convergence properties of the proposed technique were tested. They show dual-population scheme performs similarly to the regular genetic algorithms while achieves better results than the present hardware-oriented genetic algorithm models.

Automatic evaluation of skin histopathological images for melanocytic features

Conference: Medical Imaging: Digital Pathology

Authors: M. Koosha, P. Hoseini, M. Nicolescu, Z. Safaei Naraghi

Year: 2017

Link: SPIE or pdf

Abstract:

Successfully detecting melanocyte cells in the skin epidermis has great significance in skin histopathology. Because of the existence of cells with similar appearance to melanocytes in hematoxylin and eosin (HE) images of the epidermis, detecting melanocytes becomes a challenging task. This paper proposes a novel technique for the detection of melanocytes in HE images of the epidermis, based on the melanocyte color features, in the HSI color domain. Initially, an effective soft morphological filter is applied to the HE images in the HSI color domain to remove noise. Then a novel threshold-based technique is applied to distinguish the candidate melanocytes’ nuclei. Similarly, the method is applied to find the candidate surrounding halos of the melanocytes. The candidate nuclei are associated with their surrounding halos using the suggested logical and statistical inferences. Finally, a fuzzy inference system is proposed, based on the HSI color information of a typical melanocyte in the epidermis, to calculate the similarity ratio of each candidate cell to a melanocyte. As our review on the literature shows, this is the first method evaluating epidermis cells for melanocyte similarity ratio. Experimental results on various images with different zooming factors show that the proposed method improves the results of previous works.

A novel evolutionary algorithm for block-based neural network training

Conference: Iranian Conference on Pattern Recognition and Image Analysis

Authors: A. Niknam, P. Hoseini, B. Mashoufi, A. Khoei

Year: 2013

Link: IEEE or pdf

Abstract:

A novel evolutionary algorithm with fixed genetic parameters rate have presented for block-based neural network (BbNN) training. This algorithm can be used in BbNN training which faces complicated problems such as simulation of equations, classification of signals, image processing and implementation of logic gates and so on. The fixed structure of our specific BbNN allows us to implement the trained network by a fixed circuit rather than utilizing a reconfigurable hardware which is usually employed in conventional designs. Avoiding the reconfigurable hardware leads to lower power consumption and chip area. All simulations are performed in MATLAB software.

Fast and flexible genetic algorithm processor

Conference: International Conference on Electronics, Circuits, and Systems

Authors: P. Hoseini, A. Khoei, K. Hadidi, S. Moshfe

Year: 2011

Link: IEEE or pdf

Abstract:

In this paper a generic genetic algorithm processor (GAP) with high flexibility in parameter tuning is introduced. The proposed processor utilizes pipeline structure to have low processing time. In order to further increase in the speed, genetic population has been duplicated, one for replacement stage of genetic algorithm (GA) and another for selection phase. Additionally, parallel processing method in the selection stage boosts GA processor's speed. The proposed GA has been designed so that it can work in online controlling circumstances. It supports for constraints in search space and changing environments. Also, a large bit number of chromosomes can be achieved by connecting the proposed 32-bit processors to work as one n-bit chip. Ability to work with two fitness function chips, supporting pipelined fitness functions, and capability of distributed processing are other factors that increase the speed in our design.

Design of a programmable analog CMOS rational-powered membership function generator in current mode approach

Conference: International Conference on Electronics, Circuits, and Systems

Authors: S. Moshfe, A. Khoei, K. Hadidi, P. Hoseini

Year: 2011

Link: IEEE or pdf

Abstract:

This paper presents a novel structure to implement the rational-powered membership functions (RPMF), that are the extended forms of triangular/trapezoidal membership functions and those functions which are generated by applying linguistic hedges. Proposed method is based on the approximation of the function (x a ) by the functions square and square/rooter which are simply implemented in current mode analog approach based on trans-linear principle which leads us to excel in simplicity, high accuracy, and less hardware usage, and therefore lower chip area and less power consumption. Designed circuit were simulated by HSPICE simulator with level 49 parameters (BSIM3v3) and the simulation results show the average power consumption is 800 μwatt, while the RMS error is 1.25%. Finally, layout of the circuit is presented.

High speed area reduced 64-bit static hybrid carry-lookahead/carry-select adder

Conference: International Conference on Electronics, Circuits, and Systems

Authors: H. G. Tamar, A. G. Tamar, K. Hadidi, A. Khoei, P. Hoseini

Year: 2011

Link: IEEE or pdf

Abstract:

In carry-select adders (CSAs), using a single ripple carry adder and a first zero finder (FZF) circuit instead of dual ripple carry adder has an impressive impact on reduction of number of transistors and so power consumption of adder. On the other hand, combination of CSA and carry-lookahead adder (CLA) improves speed of this adder. In this paper a 64-bit static adder with structure of a hybrid CLA/CSA is presented. This adder operates with low power and occupies lower area in comparison to conventional CSA circuit due to using a first zero finder circuit. Besides by three basic changes in the critical path of adder, speed is improved considerably; First of all we used a high speed compact CLA as partial adder in each block, then a block carry generator (BCG) circuit is used for faster carry propagation and finally we replaced multiplexer gate with a XNOR gate. This circuit is implemented in TSMC 0.18μm CMOS technology at 1.8V power supply. Critical path delay of this adder decreased to 592ps, equivalent to 7.6 FO4 (fanout-of-4) inverter delays.

Hybrid ant colony optimization, genetic algorithm, and simulated annealing for image contrast enhancement

Conference: IEEE Congress on Evolutionary Computation

Authors: P. Hoseini, M. G. Shayesteh

Year: 2010

Link: IEEE or pdf

Abstract:

In this paper, we propose a hybrid algorithm including Genetic Algorithm (GA), Ant Colony Optimization (ACO), and Simulated Annealing (SA) metaheuristics for increasing the contrast of images. In this way, the contrast enhancement is obtained by globally transformation of the input intensities. ACO is used to generate the transfer functions which map the input intensities to the output intensities. SA as a local search method is utilized to modify the transfer functions generated by ACO. GA has the responsibility of evolutionary process of ants' characteristics. The results indicate that the new method performs better than the previously presented methods from the subjective and objective viewpoints.

Circuit design of voltage mode center of gravity defuzzifier in CMOS process

Conference: International Conference on Electronic Devices, Systems and Applications

Authors: P. Hoseini, A. Khoei, K. Hadidi

Year: 2010

Link: IEEE or pdf

Abstract:

In this paper a voltage input-output center of gravity (COG) defuzzifier circuit is designed without using divider in CMOS 0.35 μm process. We have used transconductance amplifier (TCA) structure as a multiplier with voltage-input - current-output for implementation of defuzzifier by exploiting of voltage follower aggregation principle. Good results have been obtained from the point of accuracy and speed.

Circuit design of weighted order statistics filter based on neural network in CMOS process

Conference: International Conference on Electronic Devices, Systems and Applications

Authors: P. Hoseini, B. Mashoufi

Year: 2010

Link: IEEE or pdf

Abstract:

In this paper a circuit of order statistics filter is proposed which is programmable as a result of weighting action. In this design a sign function circuit is offered based on the neural network structure. One of the main applications of the proposed filter is in the image processing and for this reason all examples and operations are assumed for images. The circuit has been designed in CMOS 0.35 μm process and good results have been achieved.

ORCID
Mendeley
Google Scholar
LinkedIn
GitHub
ResearchGate
YouTube
Email