Frank Boland

Stochastic convergence of the LMS algorithm in adaptive systems

Abstract This paper presents an analysis of LMS-driven adaptive filters based on stochastic stability theory. Previous work in this area is critically reviewed, and the nature of convergence of sequences of random variables and of stochastic stability are discussed. In applying a theorem of Kushner to the LMS adaptive filter, it is found that the system is initially exponentially stable with the excess mean square error tending to zero. However, eventually a point in the convergence process is reached beyond which further convergence cannot be guaranteed. With the system in a state of dynamic equilibrium, an expression is derived for an upper bound to the residual mean square error. The compromise involved in balancing the conflicting requirements of fast rate of convergence and low steady-state residual error is examined and the effect of inadequate filter length is considered. The theoretical assertions are supported by a set of computer simulation experiments.

A new phase extraction algorithm for phase profilometry

Abstract. This paper describes a new phase extraction algorithm for 3D optical profilometry based on the projection of a periodic light pattern and phase measurement (phase profilometry). The algorithm uses a square wave to demodulate phase and moving averages and comb-shaped filters to extract the phase information from low frequency. The proposed algorithm is compared with the two major profilometry techniques, namely Fourier domain profilometry and signal domain profilometry based on FIR low-pass filtering. Comparison is focused on adaptiveness to changes of the pattern frequency, ability to deal with nonuniform surfaces and computational complexity. Adaptiveness analysis is carried out by means of simulations. The issue of nonuniform surfaces is discussed on the basis of experimental results obtained from application of phase profilometry to on-line 3D printed circuit board inspection. With regard to complexity, theoretical estimates are verified by means of actual computation time measurements.

Solder-paste inspection by structured light methods based on phase measurement

Solder paste inspection, i.e. reconstruction of the 3D shape of the solder paste printed on surface mount technology component pads, is a major inspection task in the printed circuit board manufacturing process. We have investigated the use of machine vision methods based on structured light and phase measurement for on-line solder paste inspection. With these methods a periodic light pattern is projected onto the object and imaged by a camera from a different viewpoint. Since the imaged pattern is phase-modulated according to the topography of the object, suitable phase- extraction techniques enable reconstruction of the 3D shape. In a previous paper we have shown that the Fourier-based phase-extraction technique allows recovery of the 3D shape of solder paste on pads but is too slow for effective on- line inspection. In this paper we describe a new profilometry technique which uses a fast demodulation scheme and simple DSP operators such as comb filters and moving averages. Experimental results show that the new technique is equivalent to the Fourier technique with respect to reconstruction of the 3D shape of solder paste and computation times measurements indicate that it is roughly 20 times faster.

Localization accuracy of advanced spatialisation techniques in small concert halls

A comparison of several spatialization systems is presented in terms of their localization accuracy under the nonideal listening conditions found in typical small concert halls. Of interest is the effect of real reverberant conditions, noncentral listening positions, and nonsymmetrical speaker arrays on source localization. The data are presented by comparison of empirical binaural measurements and perceptual listening tests carried out using Ambisonics, VBAP, SPAT∼, and Delta Stereophony systems. The listening tests are conducted by comparing the localization of phantom sources generated by the spatialization systems to sources generated by reference loudspeakers. The reference and phantom sources are presented at front, side, and back locations about a nine‐listener audience, and the systems are tested in a random order with a calibrated 16‐loudspeaker array situated nonsymmetrically around the audience area. The binaural recordings are compared to the subjective measurements of localization accuracy through the interaural level and time difference cues for each source position.

Analysis of the stroboscopic waveform mode in scanning electron microscopy

An analysis is presented of the stroboscopic waveform mode for electron beam testing of integrated circuits. The significance of the analysis is the formulation of the relationships between the duration and phase increment of the electron beam and the bandwidth of the signal to be measured. The results provide a strategy for obtaining efficient settings of the system parameters. The analysis is supported by an examination of experimental data.

Using cooperative learning to enhance critical reflection

Over the past two decades, one of the key changes in Engineering curricula has been in the emphasis placed on the development of personal transferable skills. Employers also profess to seek these competencies when making recruiting decisions. Such skills include the ability to learn through collaboration and through meaningful critical reflection on ones own performance, both as an individual and in a group setting. However, there is still much debate on the most effective mechanisms for acquiring such skills. Engineering students can be extremely resistant to the notion of individual or collective reflection and often find it difficult to function successfully in a group setting. This paper reports on how cooperative learning has been integrated into a large-scale, practical freshman Engineering laboratory module, and explores the student experiences of being encouraged to engage in meaningful reflection. In particular, it details the presentation modalities and phenomena that influence the engineering students attitudes toward cooperative learning and reflection.

Voice activity detection

A method for detecting voice activity comprises receiving audio signals on a plurality of channels and processing the audio signals on the channels e.g. by beamformers 200 to improve the signal-to-noise ratio thereof. The processed audio signals on each channel are then fed to associated voice activity detection algorithms 202 and further processed. A voice or silence determination is then rendered by decision logic 204 based on at least the output of the voice activity detection algorithms. The method is useful in talker localization systems e.g. for teleconferencing.

Depth perception of audio sources in stereo 3D environments

In this paper we undertook perceptual experiments to determine the allowed differences in depth between audio and visual stimuli in stereoscopic-3D environments while being perceived as congruent. We also investigated whether the nature of the environment and stimuli affects the perception of congruence. This was achieved by creating an audio-visual environment consisting of a photorealistic visual environment captured by a camera under orthostereoscopic conditions and a virtual audio environment generated by measuring the acoustic properties of the real environment. The visual environment consisted of a room with a loudspeaker or person forming the visual stimulus and was presented to the viewer using a passive stereoscopic display. Pink noise samples and female speech were used as audio stimuli which were presented over headphones using binaural renderings. The stimuli were generated at different depths from the viewer and the viewer was asked to determine whether the audio stimulus was nearer, further away or at the same depth as the visual stimulus. From our experiments it is shown that there is a significant range of depth differences for which audio and visual stimuli are perceived as congruent. Furthermore, this range increases as the depth of the visual stimulus increases.

Voxel-based Viterbi Active Speaker Tracking (V-VAST) with best view selection for video lecture post-production

An automated system is presented for reducing a multi-view lecture recording into a single view video containing a best view summary of active speakers. The system uses skin color detection and voxel-based analysis in locating likely speaker locations. Using time-delay estimates from multiple microphones, speech activity is analyzed for each speaker position. The Viterbi algorithm is then used to estimate a track of the active speaker which maximizes the observed speech activity. This novel approach is termed Voxel-based Viterbi Active Speaker Tracking (V-VAST) and is shown to track speakers with an accuracy of 0.23m. Using the tracking information, the system then extracts from the available camera views the most frontal face view of the active speaker to display.

Automated registration of low and high resolution atomic force microscopy images using scale invariant features

This paper introduces a method for registering scans acquired by Atomic Force Microscopy (AFM). Due to compromises between scan size, resolution, and scan rate, high resolution data is only attainable in a very limited field of view. The proposed method uses a sparse set of feature matches between the low and high resolution AFM scans and maps them onto a common coordinate system. This can provide a wider field of view of the sample and give context to the regions where high resolution AFM data has been obtained. The algorithm employs a robust approach overcoming complications due to temporal sample changes and sample drift of the AFM system which becomes significant at higher-resolutions. To our knowledge, this is the first approach for automatic high resolution AFM image registration. Experimental results show the correctness and robustness of our approach and shows that the estimated transforms can be used to deduce plausible measures of sample drift.