No abstract

This paper deals with automatic taxa identification based on machine learning methods. The aim is therefore to automatically classify diatoms, in terms of pattern recognition terminology. Diatoms are a kind of algae microorganism with high biodiversity at the species level, which are useful for water quality assessment. The most relevant features for diatom description and classification have been selected using an extensive dataset of 80 taxa with a minimum of 100 samples/taxon augmented to 300 samples/taxon. In addition to published morphological, statistical and textural descriptors, a new textural descriptor, Local Binary Patterns (LBP), to characterize the diatom's valves, and a log Gabor implementation not tested before for this purpose are introduced in this paper. Results show an overall accuracy of 98.11% using bagging decision trees and combinations of descriptors. Finally, some phycological features of diatoms that are still difficult to integrate in computer systems are discussed for future work.

Passive time reversal has aroused considerable interest in underwater communications as a computationally inexpensive means of mitigating the intersymbol interference introduced by the channel using a receiver array. In this paper the basic technique is extended by adaptively weighting sensor contributions to partially compensate for degraded focusing due to mismatch between the assumed and actual medium impulse responses. Two algorithms are proposed, one of which restores constructive interference between sensors, and the other one minimizes the output residual as in widely used equalization schemes. These are compared with plain time reversal and variants that employ postequalization and channel tracking. They are shown to improve the residual error and temporal stability of basic time reversal with very little added complexity. Results are presented for data collected in a passive time-reversal experiment that was conducted during the MREA'04 sea trial. In that experiment a single acoustic projector generated a 2 / 4-PSK ͑phase-shift keyed͒ stream at 200/ 400 baud, modulated at 3.6 kHz, and received at a range of about 2 km on a sparse vertical array with eight hydrophones. The data were found to exhibit significant Doppler scaling, and a resampling-based preprocessing method is also proposed here to compensate for that scaling.

This paper proposes a vector sensor measurement model and the related Bartlett estimator based on particle velocity measurements for generic parameter estimation, illustrating the advantages of the Vector Sensor Array ͑VSA͒. A reliable estimate of the seabed properties such as sediment compressional speed, density and compressional attenuation based on matched-field inversion ͑MFI͒ techniques can be achieved using a small aperture VSA. It is shown that VSAs improve the resolution of seabed parameter estimation when compared with pressure sensor arrays with the same number of sensors. The data considered herein was acquired by a four-element VSA in the 8-14 kHz band, during the Makai Experiment in 2005. The results obtained with the MFI technique are compared with those obtained with a method proposed by C. Harrison, which determines the bottom reflection loss as the ratio between the upward and downward beam responses. The results show a good agreement and are in line with the historical information for the area. The particle velocity information provided by the VSA increases significantly the resolution of seabed parameter estimation and in some cases reliable results are obtained using only the vertical component of the particle velocity.

Abstract. An essential and indispensable component of automated microscopy framework is the automatic focusing system, which determines the in-focus position of a given field of view by searching the maximum value of a focusing function over a range of z-axis positions. The focus function and its computation time are crucial to the accuracy and efficiency of the system. Sixteen focusing algorithms were analyzed for histological and histopathological images. In terms of accuracy, results have shown an overall high performance by most of the methods. However, we included in the evaluation study other criteria such as computational cost and focusing curve shape which are crucial for real-time applications and were used to highlight the best practices.