Local Maxima Detection Research Articles

Automated analysis of Fourier images for semiconductor wafer monitoring

A method for automating the analysis of images generated from the Fourier imaging (FI) process is described. Fourier imaging is used to observe the evolving topography of wafers as they are being etched in a reactive ion etcher. The focus of the research described in this paper is the segmentation and analysis of images generated from FI. A brief overview of the theory of FI is presented. Segmentation of a Fourier image is done in three steps; local maxima detection, Fourier component extent identification, and boundary tracking. Quantification descriptions for each component are then generated from the segmented image. The algorithm presented demonstrates how it is possible to obtain subpixel resolution automatically in image analysis to collect data that may be used to compute critical dimensions of the microtopography of the wafer substrate.

Dynamic two-strip algorithm in curve fitting

In this paper, a new dynamic strip algorithm for fitting a curve with lines is presented. The algorithm first finds the best fitted left-hand side and right-hand side strips at each point on the curve. A figure of merit is computed based on the fitting results. Then, a local maximum detection process is applied to pick out points of high curvature. The approximated curve is one with the high curvature points connected by straight lines. The global structure of a shape, which is insensitive to a scale and orientation changes, is obtained by merging the results from several runs of the algorithm with different minimum strip widths.

Spreading activation layers, visual saccades, and invariant representations for neural pattern recognition systems

This paper shows how a simple spreading activation network in the form of two-dimensional (2D) diffusion followed by local maximum detection can quickly perform a large number of early vision tasks, for example, feature extraction, feature clustering, feature-centroid determination, and boundary gap completion, all on multiple scales. The results of the spreading activation process can be used to facilitate 2D object learning and recognition from silhouettes by generating representations from bottom-up fixation cues which are invariant to translation, orientation, and scale. In addition, the proposed network suggests a possible approach to longrange apparent motion correspondence and multiscale object decomposition. The theory is described and implementation examples are presented.

Computer analysis of ballistocardiograms

Analysis of ballistocardiograms using a small scientific digital computer is described. Two steps in processing the body accelerations and concurrent carotid pulses are employed. Desired sequences of heartbeats are selected from data recorded on FM analog magnetic tape. These are digitized and stored permanently on Linc digital magnetic tape. A “scroll-edit” program is then employed to identify all heartbeats to be analyzed. This program utilizes a cathode ray storage display unit for visual identification of each heartbeat, and teletypewriter controlled salvage of beginning and ending epochs. The digital tape generated by this procedure is analyzed by the record reading algorithm in the next step. This program works from the carotid pulse data with detection of the anacrotic limb from smoothed first differences at five millisecond intervals. Timing of the ballistocardiographic waves and measurement of their amplitudes is effected by detection of local maxima and minima of smoothed and leveled heartbeats. Correlations of wave shapes of subdivisions of the cardiac cycle are made with those from an ideally normal heartbeat. Summary calculations of each sequence of waves give average waveform and variability in a variety of printer and plotter options.