Recent Patents on Signal Processing (Discontinued) - Volume 4, Issue 2, 2014

Several approaches for optimization of fluorescence lifetime imaging microscopy (FLIM) system have been recently suggested. This paper discusses the influences of photons losses on the optimization of FLIM systems based time-correlated single photon counting (TCSPC) technique, considering the limitations associated with detecting the required amount of photons by the system. The fluorescence intensity (FI) and fluorescence lifetime (FLT) were measured in different operating regimes of the imaging system. The relation between parameters such as excitation power, detector gain, laser repetition rate, is also analyzed. Based on data acquisition limitations of typical TCSPC systems, we discuss the considerations for choosing the correct system parameters, which would most influence the accuracy of FLIM experiments. A simple scheme for patent optimization of FLIM systems for different types of fluorescent samples is finally suggested.

The scientific and accurate prediction for state of bearing is the key to ensure its safe operation. A rotating bearing monitoring system was presented in U.S. Patent 7606673 and a bearing condition monitoring apparatus was presented in U.S. Patent 8229682, however, the system or apparatus lacks state prediction function of bearing. State prediction of bearing based on relevance vector regression algorithm with RBF kernel is proposed in this paper. Kurtosis of bearing vibration signal can excellently reflect the state of bearing, so the future state of bearing can be excellently reflected by predicting the kurtosis of bearing vibration signal. Thus, kurtosis prediction of bearing vibration signal based on relevance vector regression algorithm with RBF kernel is studied. Finally, the experiments are adopted to demonstrate the feasibility of the proposed method for state prediction of bearing.

The THREAD Act (Transportation Recall Enhancement, Accountability and Documentation) mandated the use of a suitable tyre pressure monitoring system (TPMS) technology in all light motor vehicles under 5 tons. In the United States, as of 2008 and the European Union, as of November 1, 2012, all new passenger car models released must be equipped with a TPMS. This would alert drivers of under-inflation events. Many countries followed the adoption of TPMS into vehicles. The existing systems depend on pressure sensors strapped on the rim of the tyre. These sensors read the pressure information inside the tyre and transmit it to the receiver in the car. Some systems depend on wheel speed data from the wheel speed sensor. A difference in wheel speed would trigger an alarm based on the algorithm implemented. This paper proposes a new method to monitor tyre pressure by utilising the machine learning approach. Vertical vibrations are extracted from a wheel hub of a moving vehicle using an accelerometer and are classified using machine learning techniques. Statistical features are used to represent the data in the signal. The logistic model tree (LMT) was used as the classifier and attained an accuracy of 92.5% with 10 fold cross validation and 98.5 when tested.

Block matching motion estimation is a popular method in developing video coding applications. The use of fixed pattern prevents the motion estimation algorithms from locating the actual position of the global distortion minimum. Hence, to address the problem of local optima in motion estimation, it can be viewed as an optimization process to find the best matching block with reduced number of search points which is solved by the PSO technique. Due to the center biased nature of the videos, the proposed approach uses an initial pattern to speed up the convergence of the algorithm. To further improve the performance an initial search center prediction and early termination are integrated with pattern based PSO. Simulation results show that the proposed approach has significantly reduced the number of search points when compared to other fixed pattern fast block matching algorithms, without degradation in quality. In this paper we have considered several recent patents “Motion estimation methods and systems in video encoding for battery-powered appliances”, “Motion estimation for mobile device user interaction”, “Simple next search position selection for motion estimation iterative search”.

Chemometrics, a relatively young area of analytical chemistry involves extracting meaningful information from experimental data by utilizing mathematical, statistical, and computational methods with the ultimate goal of evaluating and interpreting analytical data or signals. In this article, we provide a comparison of the performance of various chemometrics techniques using data obtained from the colored complexes formed between the ligand 4-(2-pyridylazo)resorcinol (PAR) and zinc, copper, nickel, manganese, and cobalt metals. Using a reduced Box-Behnken experimental design, we developed calibration models (n=15) and compared the performance of partial least squares (PLS), principal component regression (PCR), multiple linear regression (MLR) (K-matrix), and ridge regression (RR) chemometric techniques in an independent test set (n=7). Our results show that PLS mostly outperformed the other techniques in almost all metal components in the test set. This article provides an overview of various multivariate regression techniques used in chemometrics and provides a comparison of their performance for multi-analyte detection. A discussion of some recently developed multivariate regression techniques in chemometrics as well as challenges and future directions in this field are also discussed.

Mostly pure tone fault of loudspeakers in the world is detected by human hearing. Obviously the accuracy could not be guaranteed due to subjectivity and it is easy to cause auditory fatigue. Based on the characteristics of loudspeaker’s pure tone detection, we propose that the response signal of frequency sweep can be converted into two-dimension timefrequency image signal to enhance the characteristics of fault information through wavelet packet transform. Then time-frequency images are pretreated into contours by binarization and edge extraction. The boxcounting dimensions of time-frequency image by image fractal method is proposed and regarded as the fault characteristics for loudspeaker detection. Through the verification of on-line experiments in workshop, the fractal dimension which regarded as complexity of the time-frequency image contours can be the feature for failure determination, and the fault identification accuracy rate can reach 95%. It fully meets the requirements of loudspeakers fault detection on-line and better than other recent patents.