Evaluation of Speaker Normalization Methods for Vowel Recognition Using Fuzzy ARTMAP and K-NN

A procedure that uses fuzzy ARTMAP and K-Nearest Neighbor (K-NN) categorizers to evaluate intrinsic and extrinsic speaker normalization methods is described. Each classifier is trained on preprocessed, or normalized, vowel tokens from about 30% of the speakers of the Peterson-Barney database, then tested on data from the remaining speakers. Intrinsic normalization methods included one nonscaled, four psychophysical scales (bark, bark with end-correction, mel, ERB), and three log scales, each tested on four different combinations of the fundamental (Fo) and the formants (F1 , F2, F3). For each scale and frequency combination, four extrinsic speaker adaptation schemes were tested: centroid subtraction across all frequencies (CS), centroid subtraction for each frequency (CSi), linear scale (LS), and linear transformation (LT). A total of 32 intrinsic and 128 extrinsic methods were thus compared. Fuzzy ARTMAP and K-NN showed similar trends, with K-NN performing somewhat better and fuzzy ARTMAP requiring about 1/10 as much memory. The optimal intrinsic normalization method was bark scale, or bark with end-correction, using the differences between all frequencies (Diff All). The order of performance for the extrinsic methods was LT, CSi, LS, and CS, with fuzzy AHTMAP performing best using bark scale with Diff All; and K-NN choosing psychophysical measures for all except CSi.

EVALUATION OF HARMONIC MOTION ELASTOGRAPHY AND ACOUSTO-­OPTIC IMAGING FOR MONITORING LESION FORMATION BY HIGH INTENSITY FOCUSED ULTRASOUND

Malignant or benign tumors may be ablated with high‐intensity focused ultrasound (HIFU). This technique, known as focused ultrasound surgery (FUS), has been actively investigated for decades, but slow to be implemented and difficult to control due to lack of real‐time feedback during ablation. Two methods of imaging and monitoring HIFU lesions during formation were implemented simultaneously, in order to investigate the efficacy of each and to increase confidence in the detection of the lesion. The first, Acousto‐Optic Imaging (AOI) detects the increasing optical absorption and scattering in the lesion. The intensity of a diffuse optical field in illuminated tissue is mapped at the spatial resolution of an ultrasound focal spot, using the acousto‐optic effect. The second, Harmonic Motion Imaging (HMI), detects the changing stiffness in the lesion. The HIFU beam is modulated to force oscillatory motion in the tissue, and the amplitude of this motion, measured by ultrasound pulse‐echo techniques, is influenced by the stiffness. Experiments were performed on store‐bought chicken breast and freshly slaughtered bovine liver. The AOI results correlated with the onset and relative size of forming lesions much better than prior knowledge of the HIFU power and duration. For HMI, a significant artifact was discovered due to acoustic nonlinearity. The artifact was mitigated by adjusting the phase of the HIFU and imaging pulses. A more detailed model of the HMI process than previously published was made using finite element analysis. The model showed that the amplitude of harmonic motion was primarily affected by increases in acoustic attenuation and stiffness as the lesion formed and the interaction of these effects was complex and often counteracted each other. Further biological variability in tissue properties meant that changes in motion were masked by sample‐to‐sample variation. The HMI experiments predicted lesion formation in only about a quarter of the lesions made. In simultaneous AOI/HMI experiments it appeared that AOI was a more robust method for lesion detection.

Indexing Methods for Efficient Multiclass Recognition

Many real world image analysis problems, such as face recognition and hand pose estimation, involve recognizing a large number of classes of objects or shapes. Large margin methods, such as AdaBoost and Support Vector Machines (SVMs), often provide competitive accuracy rates, but at the cost of evaluating a large number of binary classifiers, thus making it difficult to apply such methods when thousands or millions of classes need to be recognized. This thesis proposes a filter-and-refine framework, whereby, given a test pattern, a small number of candidate classes can be identified efficiently at the filter step, and computationally expensive large margin classifiers are used to evaluate these candidates at the refine step. Two different filtering methods are proposed, ClassMap and OVA-VS (One-vs.-All classification using Vector Search). ClassMap is an embedding-based method, works for both boosted classifiers and SVMs, and tends to map the patterns and their associated classes close to each other in a vector space. OVA-VS maps OVA classifiers and test patterns to vectors based on the weights and outputs of weak classifiers of the boosting scheme. At runtime, finding the strongest-responding OVA classifier becomes a classical vector search problem, where well-known methods can be used to gain efficiency. In our experiments, the proposed methods achieve significant speed-ups, in some cases up to two orders of magnitude, compared to exhaustive evaluation of all OVA classifiers. This was achieved in hand pose recognition and face recognition systems where the number of classes ranges from 535 to 48,600.