Data-Augmentation for Reducing Dataset Bias in Person Re-identification

In this paper we explore ways to address the issue of dataset bias in person re-identification by using data augmentation to increase the variability of the available datasets, and we introduce a novel data augmentation method for re-identification based on changing the image background. We show that use of data augmentation can improve the cross-dataset generalisation of convolutional network based re-identification systems, and that changing the image background yields further improvements.

Slot width augmentation in a slotted-wall transonic linear cascade wind tunnel

Transonic tests in linear cascade wind tunnels can suffer
from significant test section boundary interference effects in pitch. A slotted tailboard has been designed and optimised with an in-house Euler numerical method to reduce such ef- fects. Wind tunnel measurements on an overspeed Mach 1.27 discharge from a Rolls-Royce T2 cascade, featuring strong end-wall shock-induced interference, showed a 77% reduction in the flow pitchwise periodicity error with the optimised tail- board, with respect to the baseline open-jet cascade flow. Two-dimensional Euler predictions were also cross-validated against a three-dimensional Reynolds averaged computation, to explore the three-dimensionality of the discharge

Selecting Robust Strategies in RTS Games via Concurrent Plan Augmentation

Thesis (Master's)--University of Washington, 2015

Dual augmentation for aerobic bioremediation of MTBE and TCE pollution in heavy metal-contaminated soil

In this work we isolated from soil and characterized several bacterial strains capable of either resisting high concentrations of heavy metals (Cd2+ or Hg2+ or Pb2+) or degrading the common soil and groundwater pollutants MTBE (methyl-tertbutyl ether) or TCE (trichloroethylene). We then used soil microcosms exposed to MTBE (50 mg/l) or TCE (50 mg/l) in the presence of one heavy metal (Cd 10 ppm or Hg 5 ppm or Pb 50 or 100 ppm) and two bacterial isolates at a time, a degrader plus a metalresistant strain. Some of these two-membered consortia showed degradation efficiencies well higher (49–182% higher) than those expected under the conditions employed, demonstrating the occurrence of a synergetic relationship between the strains used. Our results show the efficacy of the dual augmentation strategy for MTBE and TCE bioremediation in the presence of heavy metals.

Augmentation cystoplasty using pedicled and de-epithelialized gastric patches in the mini-pig model.

PURPOSE: The most common methods of bladder augmentation are gastrocystoplasty and enterocystoplasty. Gastrocystoplasty is advantageous due to minimal mucous secretion and a well developed muscular wall as well as good urodynamic properties of the patch. However, the permanent contact of urine with the gastric mucosa is not free of complications. We report the urodynamic, macroscopic and histological outcomes of a pedicled de-epithelialized gastric patch incorporated in the bladder. We compared the results to those of our previous study, which sought to analyze these techniques of patch coverage using sigmoid patches. MATERIALS AND METHODS: We performed 20 augmentation cystoplasties in the mini-pig model using a pedicled de-epithelialized gastric patch and 5 techniques of patch coverage. RESULTS: Three months after surgery all bladders had an increase in volume except those in which the auto-augmentation technique was used. However, all gastric patches were smaller compared to preoperative size. Many had irregular fibrosed inner surfaces and histological evaluation revealed a fibrosed newly formed submucosal layer with a complete urothelial coverage in every patch. No gastric mucosal remnant was found. CONCLUSIONS: De-epithelialized gastrocystoplasty is an attractive procedure that can increase bladder capacity as well as provide a complete urothelial lining without mucosal remnants. However, the success of this procedure seems to be limited by increased morbidity and fibrotic changes, and decreased surface of the patch.

Les 5 milliards sont là ! ! ! sans emprunt ni augmentation des charges publiques par la seule transformation de l'impôt... 2e édition

Comprend : Préambule