The visual navigation system of a UAV is a complex embedded device designed to modify the path of the platform depending on objects or events detected on the ground. In the visual field of the autopilot these events c...
详细信息
The visual navigation system of a UAV is a complex embedded device designed to modify the path of the platform depending on objects or events detected on the ground. In the visual field of the autopilot these events could be formalized as specific space-time signatures. Processing all pixels captured by the on-board camera(s) in real time with high frame rate needs huge computational effort that is often unnecessary. An adequate computational strategy would focus on the interesting locations only as in the visual system of various species. In this article we describe an automatic focusing mechanism relying on optical flow calculation for detecting moving objects on the ground, thus efficiently separating the motion of interest from ego-motion of the platform.
We have developed a novel type of correlators named searching correlator (SC) (Tokes et al., 2003) using bacteriorhodopsin (BR) as a dynamic rewritable and multiply readable holographic material. Our POAC (programmabl...
详细信息
We have developed a novel type of correlators named searching correlator (SC) (Tokes et al., 2003) using bacteriorhodopsin (BR) as a dynamic rewritable and multiply readable holographic material. Our POAC (programmable optical array computer) model can perform highly parallel optical correlations and feedforward CNN operations. However, to keep pace with digital technologies (Moore's law); newer, more effective materials are needed. The sensitivity, diffraction efficiency and cycle time are the most critical parameters of the dynamic holographic materials to be used in 2D optical processors. Only highly engineered, composite media, such as multiple quantum well optically addressable spatial light modulator (MQW-OASLM), liquid crystal embedded polymers or photonic crystal devices will be able to satisfy all of the requirements. However, these devices could not reach yet the stage of a commercial product. That is why we are working on a new approach for realizing competitive optical CNN processor architectures by integrating a high-end CMOS camera and an LCD micro-display with an embedded FPGA processor. Embedding this PRC-OASLM, our present feed-forward-only optical CNN processor architecture can be developed to a nonlinear feedback architecture
The CNN literature contains several papers about how to use the genetic algorithms for template tuning. This paper shows a possible CNN-UM implementation of the control algorithm of the genetic algorithm in a special ...
详细信息
The CNN literature contains several papers about how to use the genetic algorithms for template tuning. This paper shows a possible CNN-UM implementation of the control algorithm of the genetic algorithm in a special fine-grained parallel version. The interesting of the analogic mapping is the different set of operators, which can be evaluated fast and efficient on the CNN-UM platform. The rational behind the spatial implementation is the possibility of focal-plane optimization
A multitarget tracking framework implemented on the Bi-i platform is presented. The demonstration applications include a target tracking system with laser actuation, an attention-selection algorithm and a laser dot de...
详细信息
A multitarget tracking framework implemented on the Bi-i platform is presented. The demonstration applications include a target tracking system with laser actuation, an attention-selection algorithm and a laser dot detection system with gaming applications [4]
systems for visual inspection, usually, are quite sensitive against variations of the optical properties of the sample surface. In contrary to the above statement, the paper presents algorithmic solutions for visual i...
详细信息
systems for visual inspection, usually, are quite sensitive against variations of the optical properties of the sample surface. In contrary to the above statement, the paper presents algorithmic solutions for visual inspection of metal objects by using a cellular wave computer, which evaluate the changes in optical properties of the inspected surface. The obtained results show the feasibility of the proposed method
This demo displays a retina exploration system. The system is designed to explore and understand the essence of the retina channels and offers a potential input transformation for retinal implants [1] The primary moti...
详细信息
This demo displays a retina exploration system. The system is designed to explore and understand the essence of the retina channels and offers a potential input transformation for retinal implants [1] The primary motivation was to deduce a simple algorithmic skeleton from neurobiological measurements [2]. The model is tuned for several different retina channels with less than one-half percent spatial-temporal error The implementation platform of our novel neuromorphic retina model is the Bi-I [3] The Bi-I computer, containing the ACE16k CNN-UM [4] and DSP chips computes our parallel multi-channel model in video real-time. The on-line control is provided by external controllers to enable changing parameters on-line. The system will be available to experimentation.
A cellular neural/nonlinear network (CNN) is any spatial arrangement of mainly locally coupled cells, where each cell is a dynamical system which has an input, an output and a state that evolves according to some pres...
详细信息
A cellular neural/nonlinear network (CNN) is any spatial arrangement of mainly locally coupled cells, where each cell is a dynamical system which has an input, an output and a state that evolves according to some prescribed dynamical laws. Since the CNN was first introduced in 1988, research in this field has developed rapidly. The goal of this tutorial is to provide participants with a snapshot of the current state of the art and research trends. It covers the broad multi-disciplinary areas of CNN research, from theoretical aspects to applications. The presenters have experience ranging from theoretical analysis of CNN dynamics, VLSI implementation and system level applications
Small world networks have a peculiar property: even though almost all of their nodes are only locally connected, the average path length between two nodes is nearly as low as that of random networks. Networks of this ...
详细信息
Small world networks have a peculiar property: even though almost all of their nodes are only locally connected, the average path length between two nodes is nearly as low as that of random networks. Networks of this type are ubiquitous in natural systems and frequently show interesting dynamics. We investigated the behavior of a new class of cellular nonlinear networks (CNNs) where all cells are locally connected to their nearest neighbors (as in a conventional CNN), but some of the nodes have long-range shortcuts as well. We give lower bounds on the number of extra links that must be added for a network to be considered as small-world and suggest an optimal topology for the added links. We also show some interesting behaviors of this network in diffusion and wave propagation phenomena.
In this paper several pattern formation phenomena generated on a prototype complex cell (second order) CNN-UM chip (CACE1K) will be demonstrated. It is shown that the basic motifs (patches, checkers, and stripes) can ...
详细信息
In this paper several pattern formation phenomena generated on a prototype complex cell (second order) CNN-UM chip (CACE1K) will be demonstrated. It is shown that the basic motifs (patches, checkers, and stripes) can be easily calculated and furthermore, the interaction of the two dynamic layers leads to new patterns that cannot be generated on a first order architecture. Throughout the paper a rich set of measured phenomena is given motivating further research toward a constructive use of pattern formation dynamics in various flow processing algorithms.
In this work we discuss feature/signature selection strategies for on-line classifier systems, implemented on a common stand-alone HW/SW vision system. In the chosen computational environment topographic and non-topog...
详细信息
In this work we discuss feature/signature selection strategies for on-line classifier systems, implemented on a common stand-alone HW/SW vision system. In the chosen computational environment topographic and non-topographic computing can be combined for the targeted task of terrain feature analysis. The topographic front-end of the system is capable of deriving more than thousand different visual signatures per second from a few dozen visual feature channels. These signatures are statistically correlated and might also be computationally inter-dependent. To develop an efficient real-time classification algorithm the dimensionality of the extracted features has to be significantly reduced, i.e. only a few signatures can be selected. In this paper we analyze different signature selection strategies for various classifiers with given performance criteria and system-level time performance constraints. Two blind signature selection techniques are described: variance maximization and a factor-based analysis. We also discuss supervised selection mechanisms: class-based statistical and decision-tree-based approaches.
暂无评论