Parameter Sensitive Detectors

Object detection can be challenging when the object class exhibits large variations. One commonly-used strategy is to first partition the space of possible object variations and then train separate classifiers for each portion. However, with continuous spaces the partitions tend to be arbitrary since there are no natural boundaries (for example, consider the continuous range of human body poses). In this paper, a new formulation is proposed, where the detectors themselves are associated with continuous parameters, and reside in a parameterized function space. There are two advantages of this strategy. First, a-priori partitioning of the parameter space is not needed; the detectors themselves are in a parameterized space. Second, the underlying parameters for object variations can be learned from training data in an unsupervised manner. In profile face detection experiments, at a fixed false alarm number of 90, our method attains a detection rate of 75% vs. 70% for the method of Viola-Jones. In hand shape detection, at a false positive rate of 0.1%, our method achieves a detection rate of 99.5% vs. 98% for partition based methods. In pedestrian detection, our method reduces the miss detection rate by a factor of three at a false positive rate of 1%, compared with the method of Dalal-Triggs.

Learning a Family of Detectors

Object detection and recognition are important problems in computer vision. The challenges of these problems come from the presence of noise, background clutter, large within class variations of the object class and limited training data. In addition, the computational complexity in the recognition process is also a concern in practice. In this thesis, we propose one approach to handle the problem of detecting an object class that exhibits large within-class variations, and a second approach to speed up the classification processes. In the first approach, we show that foreground-background classification (detection) and within-class classification of the foreground class (pose estimation) can be jointly solved with using a multiplicative form of two kernel functions. One kernel measures similarity for foreground-background classification. The other kernel accounts for latent factors that control within-class variation and implicitly enables feature sharing among foreground training samples. For applications where explicit parameterization of the within-class states is unavailable, a nonparametric formulation of the kernel can be constructed with a proper foreground distance/similarity measure. Detector training is accomplished via standard Support Vector Machine learning. The resulting detectors are tuned to specific variations in the foreground class. They also serve to evaluate hypotheses of the foreground state. When the image masks for foreground objects are provided in training, the detectors can also produce object segmentation. Methods for generating a representative sample set of detectors are proposed that can enable efficient detection and tracking. In addition, because individual detectors verify hypotheses of foreground state, they can also be incorporated in a tracking-by-detection frame work to recover foreground state in image sequences. To run the detectors efficiently at the online stage, an input-sensitive speedup strategy is proposed to select the most relevant detectors quickly. The proposed approach is tested on data sets of human hands, vehicles and human faces. On all data sets, the proposed approach achieves improved detection accuracy over the best competing approaches. In the second part of the thesis, we formulate a filter-and-refine scheme to speed up recognition processes. The binary outputs of the weak classifiers in a boosted detector are used to identify a small number of candidate foreground state hypotheses quickly via Hamming distance or weighted Hamming distance. The approach is evaluated in three applications: face recognition on the face recognition grand challenge version 2 data set, hand shape detection and parameter estimation on a hand data set, and vehicle detection and estimation of the view angle on a multi-pose vehicle data set. On all data sets, our approach is at least five times faster than simply evaluating all foreground state hypotheses with virtually no loss in classification accuracy.

Atmospheric neutrino oscillation analysis with subleading effects in Super-Kamiokande I, II, and III

We present a search for nonzero θ13 and deviations of sin2θ23 from 0.5 in the oscillations of atmospheric neutrino data from Super-Kamiokande I, II, and III. No distortions of the neutrino flux consistent with nonzero θ13 are found and both neutrino mass hierarchy hypotheses are in agreement with the data. The data are best fit at Δm2=2.1×10-3eV2, sin2θ13=0.0, and sin2θ23=0.5. In the normal (inverted) hierarchy θ13 and Δm2 are constrained at the one-dimensional 90% C.L. to sin2θ13<0.04(0.09) and 1.9(1.7)×10 -3<Δm2<2.6(2.7)×10-3eV2. The atmospheric mixing angle is within 0.407≤sin2θ23≤0.583 at 90% C.L. © 2010 The American Physical Society.

Indication of electron neutrino appearance from an accelerator-produced off-axis muon neutrino beam.

The T2K experiment observes indications of ν(μ) → ν(e) appearance in data accumulated with 1.43×10(20) protons on target. Six events pass all selection criteria at the far detector. In a three-flavor neutrino oscillation scenario with |Δm(23)(2)| = 2.4×10(-3)  eV(2), sin(2)2θ(23) = 1 and sin(2)2θ(13) = 0, the expected number of such events is 1.5±0.3(syst). Under this hypothesis, the probability to observe six or more candidate events is 7×10(-3), equivalent to 2.5σ significance. At 90% C.L., the data are consistent with 0.03(0.04) < sin(2)2θ(13) < 0.28(0.34) for δ(CP) = 0 and a normal (inverted) hierarchy.

Development of a GEM Electronic Board (GEB) for triple-GEM detectors

Developed for use with triple GEM detectors, the GEM Electronic Board (GEB) forms a crucial part of the electronics readout system being developed as part of the CMS muon upgrade program. The objective of the GEB is threefold; to provide stable powering and ground for the VFAT3 front ends, to enable high-speed communication between 24 VFAT3 front ends and an optohybrid, and to shield the GEM detector from electromagnetic interference. The paper describes the concept and design of a large-size GEB in detail, highlighting the challenges in terms of design and feasibility of this deceptively difficult system component.

The direction aftereffect is driven by adaptation of local motion detectors

The processing of motion information by the visual system can be decomposed into two general stages; point-by-point local motion extraction, followed by global motion extraction through the pooling of the local motion signals. The direction aftereVect (DAE) is a well known phenomenon in which prior adaptation to a unidirectional moving pattern results in an exaggerated perceived direction diVerence between the adapted direction and a subsequently viewed stimulus moving in a diVerent direction. The experiments in this paper sought to identify where the adaptation underlying the DAE occurs within the motion processing hierarchy. We found that the DAE exhibits interocular transfer, thus demonstrating that the underlying adapted neural mechanisms are binocularly driven and must, therefore, reside in the visual cortex. The remaining experiments measured the speed tuning of the DAE, and used the derived function to test a number of local and global models of the phenomenon. Our data provide compelling evidence that the DAE is driven by the adaptation of motion-sensitive neurons at the local-processing stage of motion encoding. This is in contrast to earlier research showing that direction repulsion, which can be viewed as a simultaneous presentation counterpart to the DAE, is a global motion process. This leads us to conclude that the DAE and direction repulsion reflect interactions between motion-sensitive neural mechanisms at different levels of the motion-processing hierarchy.

Fabrication and characterisation of high resistivity SOI substrates for monolithic high energy physics detectors

Silicon on Insulator (SOI) substrates offer a promising platform for monolithic high energy physics detectors with integrated read-out electronics and pixel diodes. This paper describes the fabrication and characterisation of specially-configured SOI substrates using improved bonded wafer ion split and grind/polish technologies. The crucial interface between the high resistivity handle silicon and the SOI buried oxide has been characterised using both pixel diodes and circular geometry MOS transistors. Pixel diode breakdown voltages were typically greater than 100V and average leakage current densities at 70 V were only 55 nA/ sq cm. MOS transistors subjected to 24 GeV proton irradiation showed an increased SOI buried oxide trapped charge of only 3.45x1011cn-2 for a dose of 2.7Mrad

Characterisation of deuterium spectra from laser driven multi-species sources by employing differentially filtered image plate detectors in Thomson spectrometers

A novel method for characterising the full spectrum of deuteron ions emitted by laser driven multi-species ion sources is discussed. The procedure is based on using differential filtering over the detector of a Thompson parabola ion spectrometer, which enables discrimination of deuterium ions from heavier ion species with the same charge-to-mass ratio (such as C^{6 +}, O^{8 +}, etc.). Commonly used Fuji Image plates were used as detectors in the spectrometer, whose absolute response to deuterium ions over a wide range of energies was calibrated by using slotted CR-39 nuclear track detectors. A typical deuterium ion spectrum diagnosed in a recent experimental campaign is presented, which was produced from a thin deuterated plastic foil target irradiated by a high power laser.