Uncertainty-Aware Gaze Tracking for Assisted Living Environments.

Effective assisted living environments must be able to infer how their occupants interact in a variety of scenarios. Gaze direction provides strong indications of how a person engages with the environment and its occupants. In this paper, we investigate the problem of gaze tracking in multi-camera assisted living environments. We propose a gaze tracking method based on predictions generated by a neural network regressor that relies only on the relative positions of facial keypoints to estimate gaze. For each gaze prediction, our regressor also provides an estimate of its own uncertainty, which is used to weigh the contribution of previously estimated gazes within a tracking framework based on an angular Kalman filter. Our gaze estimation neural network uses confidence gated units to alleviate keypoint prediction uncertainties in scenarios involving partial occlusions or unfavorable views of the subjects. We evaluate our method using videos from the MoDiPro dataset, which we acquired in a real assisted living facility, and on the publicly available MPIIFaceGaze, GazeFollow, and Gaze360 datasets. Experimental results show that our gaze estimation network outperforms sophisticated state-of-the-art methods, while additionally providing uncertainty predictions that are highly correlated with the actual angular error of the corresponding estimates. Finally, an analysis of the temporal integration performance of our method demonstrates that it generates accurate and temporally stable gaze predictions.

Alpha spending for historical versus surveillance Poisson data with CMaxSPRT.

Sequential analysis hypothesis testing is now an important tool for postmarket drug and vaccine safety surveillance. When the number of adverse events accruing in time is assumed to follow a Poisson distribution, and if the baseline Poisson rate is assessed only with uncertainty, the conditional maximized sequential probability ratio test, CMaxSPRT, is a formal solution. CMaxSPRT is based on comparing monitored data with historical matched data, and it was primarily developed under a flat signaling threshold. This paper demonstrates that CMaxSPRT can be performed under nonflat thresholds too. We pose the discussion in the light of the alpha spending approach. In addition, we offer a rule of thumb for establishing the best shape of the signaling threshold in the sense of minimizing expected time to signal and expected sample size. An example involving surveillance for adverse events after influenza vaccination is used to illustrate the method.

Image processing for identification and quantification of filamentous bacteria in in situ acquired images.

BACKGROUND: In the activated sludge process, problems of filamentous bulking and foaming can occur due to overgrowth of certain filamentous bacteria. Nowadays, these microorganisms are typically monitored by means of light microscopy, commonly combined with staining techniques. As drawbacks, these methods are susceptible to human errors, subjectivity and limited by the use of discontinuous microscopy. The in situ microscope appears as a suitable tool for continuous monitoring of filamentous bacteria, providing real-time examination, automated analysis and eliminating sampling, preparation and transport of samples. In this context, a proper image processing algorithm is proposed for automated recognition and measurement of filamentous objects. METHODS: This work introduces a method for real-time evaluation of images without any staining, phase-contrast or dilution techniques, differently from studies present in the literature. Moreover, we introduce an algorithm which estimates the total extended filament length based on geodesic distance calculation. For a period of twelve months, samples from an industrial activated sludge plant were weekly collected and imaged without any prior conditioning, replicating real environment conditions. RESULTS: Trends of filament growth rate-the most important parameter for decision making-are correctly identified. For reference images whose filaments were marked by specialists, the algorithm correctly recognized 72 % of the filaments pixels, with a false positive rate of at most 14 %. An average execution time of 0.7 s per image was achieved. CONCLUSIONS: Experiments have shown that the designed algorithm provided a suitable quantification of filaments when compared with human perception and standard methods. The algorithm's average execution time proved its suitability for being optimally mapped into a computational architecture to provide real-time monitoring.

In situ microscopy as a tool for the monitoring of filamentous bacteria: a case study in an industrial activated sludge system dominated by M. parvicella.

The present study demonstrates the application of in situ microscopy for monitoring the growth of filamentous bacteria which can induce disturbances in an industrial activated sludge process. An in situ microscope (ISM) is immersed directly into samples of activated sludge with Microthrix parvicella as dominating species. Without needing further preparatory steps, the automatic evaluation of the ISM-images generates two signals: the number of individual filaments per image (ISM-filament counting) and the total extended filament length (TEFL) per image (ISM-online TEFL). In this first version of the image-processing algorithm, closely spaced crossing filament-segments or filaments within bulk material are not detected. The signals show highly linear correlation both with the standard filament index and the TEFL. Correlations were further substantiated by comparison with real-time polymerase chain reaction (real-time PCR) measurements of M. parvicella and of the diluted sludge volume index. In this case study, in situ microscopy proved to be a suitable tool for straightforward online-monitoring of filamentous bacteria in activated sludge systems. With future adaptation of the system to different filament morphologies, including cross-linking filaments, bundles, and attached growth, the system will be applicable to other wastewater treatment plants.

Fluorescence-based Lab-on-Chip spot design for improved signal detection.

This work presents a new perspective for fluorescence signal detection using specific optics on Lab-on-Chip devices. An apparatus was designed and implemented in order to assess the performance of a fluorescence technique using different detection spot configurations, using the chip itself as a waveguide for illumination. The experiments conducted investigate the influence of the dimensions - diameter and height - of the spot on the amplitude of the detection output signal. Results show that the configuration of optical interfaces must be considered in order to improve detection output, or to be able to detect less fluorophore molecules in the spot.