Training Bayesian networks for image segmentation

We are concerned with the problem of image segmentation in which each pixel is assigned to one of a predefined finite number of classes. In Bayesian image analysis, this requires fusing together local predictions for the class labels with a prior model of segmentations. Markov Random Fields (MRFs) have been used to incorporate some of this prior knowledge, but this not entirely satisfactory as inference in MRFs is NP-hard. The multiscale quadtree model of Bouman and Shapiro (1994) is an attractive alternative, as this is a tree-structured belief network in which inference can be carried out in linear time (Pearl 1988). It is an hierarchical model where the bottom-level nodes are pixels, and higher levels correspond to downsampled versions of the image. The conditional-probability tables (CPTs) in the belief network encode the knowledge of how the levels interact. In this paper we discuss two methods of learning the CPTs given training data, using (a) maximum likelihood and the EM algorithm and (b) emphconditional maximum likelihood (CML). Segmentations obtained using networks trained by CML show a statistically-significant improvement in performance on synthetic images. We also demonstrate the methods on a real-world outdoor-scene segmentation task.

All-optical signal regeneration by temporal slicing of nonlinearly flattened optical waveform

A novel all-optical time domain regeneration technique using nonlinear pulse broadening and flattening in normal dispersion fiber and subsequent temporal slicing by an amplitude modulator (or a device performing a similar function) is proposed. Substantial suppression of the timing jitter of jitter-degraded optical signals is demonstrated using the proposed approach.

All-optical signal regeneration by temporal slicing of nonlinearly flattened optical waveform

A novel all-optical time domain regeneration technique using nonlinear pulse broadening and flattening in normal dispersion fiber and subsequent temporal slicing by an amplitude modulator (or a device performing a similar function) is proposed. Substantial suppression of the timing jitter of jitter-degraded optical signals is demonstrated using the proposed approach. © 2005 IEEE.

Objective accommodative amplitude and dynamics with the 1CU accommodative intraocular lens

PURPOSE. To compare the objective accommodative amplitude and dynamics of eyes implanted with the one-compartment-unit (1CU; HumanOptics AG, Erlangen, Germany) accommodative intraocular lenses (IOLs) with that measured subjectively. METHODS. Twenty eyes with a 1CU accommodative IOL implanted were refracted and distance and near acuity measured with a logMAR (logarithm of the minimum angle of resolution) chart. The objective accommodative stimulus-response curve for static targets between 0.17 and 4.00 D accommodative demand was measured with the SRW-5000 (Shin-Nippon Commerce Inc., Tokyo, Japan) and PowerRefractor (PlusOptiX, Nürnberg, Germany) autorefractors. Continuous objective recording of dynamic accommodation was measured with the SRW-5000, with the subject viewing a target moving from 0 to 2.50 D at 0.3 Hz through a Badal lens system. Wavefront aberrometry measures (Zywave; Bausch & Lomb, Rochester, NY) were made through undilated pupils. Subjective amplitude of accommodation was measured with the RAF (Royal Air Force accommodation and vergence measurement) rule. RESULTS. Four months after implantation best-corrected acuity was -0.01 ± 0.16 logMAR at distance and 0.60 ± 0.09 logMAR at near. Objectively, the static amplitude of accommodation was 0.72 ± 0.38 D. The average dynamic amplitude of accommodation was 0.71 ± 0.47 D, with a lag behind the target of 0.50 ± 0.48 seconds. Aberrometry showed a decrease in power of the lens-eye combination from the center to the periphery in all subjects (on average, -0.38 ± 0.28 D/mm). Subjective amplitude of accommodation was 2.24 ± 0.42 D. Two years after 1CU implantation, refractive error and distance visual acuity remained relatively stable, but near visual acuity, and the subjective and objective amplitudes of accommodation decreased. CONCLUSIONS. The objective accommodating effects of the 1CU lens appear to be limited, although patients are able to track a moving target. Subjective and objective accommodation was reduced at the 2-year follow-up. The greater subjective amplitude of accommodation is likely to result from the eye's depth of focus of and the aspheric nature of the IOL. Copyright © Association for Research in Vision and Ophthalmology.