Acropolis Educational Resources Repository
Not a member yet
31278 research outputs found
Sort by
NeuralDiff: Segmenting 3D objects that move in egocentric videos
No full textGiven a raw video sequence taken from a freely-moving camera, we study the problem of decomposing the observed 3D scene into a static background and a dynamic foreground containing the objects that move within the scene. This task is reminiscent of the classic background subtraction problem, but is significantly harder because all parts of the scene, static and dynamic, generate a large apparent motion due to the camera large viewpoint change and parallax. In particular, we consider egocentric videos and further separate the dynamic component into objects and the actor that observes and moves them. We achieve this factorization by reconstructing the video via a triple-stream neural rendering network that explains the different motions based on corresponding inductive biases. We demonstrate that our method can successfully separate the different types of motion, outperforming recent neural rendering baselines at this task, and can accurately segment the moving objects. We do so by assessing the method empirically on challenging videos from the EPIC-KITCHENS dataset which we augment with appropriate annotations to create a new benchmark for the task of dynamic object segmentation on unconstrained video sequences, for complex 3D environments
The Parthenon Sculptures. 6 short stories of separation-Family trail
No full textΜε το φυλλάδιο «Τα γλυπτά του Παρθενώνα. 6 μικρές ιστορίες χωρισμού», οι μικροί επισκέπτες ανεβαίνουν στον 3ο όροφο του Μουσείου, στην αίθουσα του Παρθενώνα. Εκεί τους περιμένουν τα πιο διάσημα, ίσως, γλυπτά του κόσμου! Όμως όχι όλα … Ένα μέρος τους βρίσκεται αλλού … Το πού, το πότε, το πώς και το γιατί θα το ανακαλύψουν μέσα από τις περιπέτειες έξι διάσημων γλυπτών, που περιέχονται στο έντυπο. Στο τέλος τα παιδιά περιγράφουν με ζωγραφιές και λέξεις τα αισθήματα που τους γεννά ο χωρισμός των γλυπτών στις ειδικές σελίδες του εντύπου και τις αφήνουν στο Γραφείο Πληροφοριών του Μουσείου.Υoung visitors will walk towards the Parthenon Gallery on the Museum’s third floor, where they will discover the, perhaps most famous, sculptures worldwide! But not all of them ... A part of these is located elsewhere ... The details behind where, when, how and why they went missing will be discovered through the adventures of six famous sculptures showcased in the pamphlet. By the end of the activity, children will have the chance to describe their thoughts and feelings about the dispersal of the sculptures, through paintings and phrases, in the special pages of the pamphlet, which they will leave behind upon their exit from the Museum, at the Information Desk
Analysis of placental arteriovenous formation reveals new insights into embryos with congenital heart defects
No full textThe placental vasculature provides the developing embryo with a circulation to deliver nutrients and dispose of waste products. However, in the mouse, the vascular components of the chorio-allantoic placenta have been largely unexplored due to a lack of well-validated molecular markers. This is required to study how these blood vessels form in development and how they are impacted by embryonic or maternal defects. Here, we employed marker analysis to characterize the arterial/arteriole and venous/venule endothelial cells (ECs) during normal mouse placental development. We reveal that placental ECs are potentially unique compared with their embryonic counterparts. We assessed embryonic markers of arterial ECs, venous ECs, and their capillary counterparts—arteriole and venule ECs. Major findings were that the arterial tree exclusively expressed Dll4, and venous vascular tree could be distinguished from the arterial tree by Endomucin (EMCN) expression levels. The relationship between the placenta and developing heart is particularly interesting. These two organs form at the same stages of embryogenesis and are well known to affect each other’s growth trajectories. However, although there are many mouse models of heart defects, these are not routinely assessed for placental defects. Using these new placental vascular markers, we reveal that mouse embryos from one model of heart defects, caused by maternal iron deficiency, also have defects in the formation of the placental arterial, but not the venous, vascular tree. Defects to the embryonic cardiovascular system can therefore have a significant impact on blood flow delivery and expansion of the placental arterial tree
Modifying the maternal microbiota alters the gut-brain metabolome and prevents emotional dysfunction in the adult offspring of obese dams [metabolomics dataset]
No full textData collected between 2019 and 2021. Data pertains to work published in Proceedings of the National Academy of Sciences of the United States of America. This dataset contains the metabolomics sum-normalised data used in the study. The name of each spreadsheet identifies the cohort the data corresponds to
Key themes for the global energy economy in 2022
No full textThis paper contains 20 short articles which outline the views of OIES research fellows on some of the key themes for the global energy economy in 2022. Starting with views on the short-term outlook for oil, electricity and gas markets, the articles move on to cover LNG supply and Russian export strategy as well as the future of the Nord Stream 2 pipeline. The majority of the articles, though, are on issues around the energy transition, looking at the key milestones ahead of COP27 but also considering many of the most important topics discussed at COP26 where action now needs to be taken. These include the Global Methane Pledge, the development of voluntary carbon markets, the prospects for hydrogen projects, the building of a business case for CCUS and the need to update country NDCs in the next 12 months. The paper also looks at regional developments in the EU, India and China, and considers a number of legislative and regulatory topics that are likely to dominate environmental policy making in the energy sector during 2022
DensePose 3D: lifting canonical surface maps of articulated objects to the third dimension
No full textWe tackle the problem of monocular 3D reconstruction of articulated objects like humans and animals. We contribute DensePose 3D, a method that can learn such reconstructions in a weakly supervised fashion from 2D image annotations only. This is in stark contrast with previous deformable reconstruction methods that use parametric models such as SMPL pre-trained on a large dataset of 3D object scans. Because it does not require 3D scans, DensePose 3D can be used for learning a wide range of articulated categories such as different animal species. The method learns, in an end-to-end fashion, a soft partition of a given category-specific 3D template mesh into rigid parts together with a monocular reconstruction network that predicts the part motions such that they reproject correctly onto 2D DensePose-like surface annotations of the object. The decomposition of the object into parts is regularized by expressing part assignments as a combination of the smooth eigenfunctions of the Laplace-Beltrami operator. We show significant improvements compared to state-of-the-art nonrigid structure-from-motion baselines on both synthetic and real data on categories of humans and animals
A non-coding RNA balancing act: miR-346-induced DNA damage is limited by the long non-coding RNA NORAD in prostate cancer
No full textBackground
miR-346 was identified as an activator of Androgen Receptor (AR) signalling that associates with DNA damage response (DDR)-linked transcripts in prostate cancer (PC). We sought to delineate the impact of miR-346 on DNA damage, and its potential as a therapeutic agent.
Methods
RNA-IP, RNA-seq, RNA-ISH, DNA fibre assays, in vivo xenograft studies and bioinformatics approaches were used alongside a novel method for amplification-free, single nucleotide-resolution genome-wide mapping of DNA breaks (INDUCE-seq).
Results
miR-346 induces rapid and extensive DNA damage in PC cells - the first report of microRNA-induced DNA damage. Mechanistically, this is achieved through transcriptional hyperactivation, R-loop formation and replication stress, leading to checkpoint activation and cell cycle arrest. miR-346 also interacts with genome-protective lncRNA NORAD to disrupt its interaction with PUM2, leading to PUM2 stabilisation and its increased turnover of DNA damage response (DDR) transcripts. Confirming clinical relevance, NORAD expression and activity strongly correlate with poor PC clinical outcomes and increased DDR in biopsy RNA-seq studies. In contrast, miR-346 is associated with improved PC survival.
INDUCE-seq reveals that miR-346-induced DSBs occur preferentially at binding sites of the most highly-transcriptionally active transcription factors in PC cells, including c-Myc, FOXA1, HOXB13, NKX3.1, and importantly, AR, resulting in target transcript downregulation. Further, RNA-seq reveals widespread miR-346 and shNORAD dysregulation of DNA damage, replication and cell cycle processes.
NORAD drives target-directed miR decay (TDMD) of miR-346 as a novel genome protection mechanism: NORAD silencing increases mature miR-346 levels by several thousand-fold, and WT but not TDMD-mutant NORAD rescues miR-346-induced DNA damage. Importantly, miR-346 sensitises PC cells to DNA-damaging drugs including PARP inhibitor and chemotherapy, and induces tumour regression as a monotherapy in vivo, indicating that targeting miR-346:NORAD balance is a valid therapeutic strategy.
Conclusions
A balancing act between miR-346 and NORAD regulates DNA damage and repair in PC. miR-346 may be particularly effective as a therapeutic in the context of decreased NORAD observed in advanced PC, and in transcriptionally-hyperactive cancer cells
Investigations on zinc isotope fractionation in breast cancer tissue using in vitro cell culture uptake-efflux experiments
No full textZinc (Zn) accumulates in breast cancer tumors compared to adjacent healthy tissue. Clinical samples of breast cancer tissue show light Zn isotopic compositions (δ66Zn) relative to healthy tissue. The underlying mechanisms causing such effects are unknown. To investigate if the isotopic discrimination observed for in vivo breast cancer tissue samples can be reproduced in vitro, we report isotopic data for Zn uptake-efflux experiments using a human breast cancer cell line. MDA-MB-231 cell line was used as a model for triple receptor negative breast cancer. We determined Zn isotope fractionation for Zn cell uptake (Δ66Znuptake) and cell efflux (Δ66Znefflux) using a drip-flow reactor to enable comparison with the in vivo environment. The MDA-MB-231 cell line analyses show Zn isotopic fractionations in an opposite direction to those observed for in vivo breast cancer tissue. Uptake of isotopically heavy Zn (Δ66Znuptake = +0.23 ± 0.05‰) is consistent with transport via Zn transporters (ZIPs), which have histidine-rich binding sites. Zinc excreted during efflux is isotopically lighter than Zn taken up by the cells (Δ66Znefflux = −0.35 ± 0.06‰). The difference in Zn isotope fractionation observed between in vitro MDA-MB-231 cell line experiments and in vivo breast tissues might be due to differences in Zn transporter levels or intercellular Zn storage (endoplasmic reticulum and/or Zn specific vesicles); stromal cells, such as fibroblasts and immune cells. Although, additional experiments using other human breast cancer cell lines (e.g., MCF-7, BT-20) with varying Zn protein characteristics are required, the results highlight differences between in vitro and in vivo Zn isotope fractionation
Supplementary data for "extended electron tails in electrostatic microinstabilities and the nonadiabatic response of passing electrons": arXiv 2108.02822
No full textSupplementary data for the article "Extended electron tails in electrostatic microinstabilities and the nonadiabatic response of passing electrons": arXiv 2108.02822. The dataset includes a readme, GS2 FORTRAN namelist input files necessary to reproduce the simulations presented in the article, as well as scripts (using a mixture of Mathematica, MATLAB, and Python) for the calculation of collisional transport coefficients that appear in the collisional theory of the studied microinstabilities
Learning inertial odometry for dynamic legged robot state estimation
No full textThis paper introduces a novel proprioceptive state estimator for legged robots based on a learned displacement measurement from IMU data. Recent research in pedestrian tracking has shown that motion can be inferred from inertial data using convolutional neural networks. A learned inertial displacement measurement can improve state estimation in challenging scenarios where leg odometry is unreliable, such as slipping and compressible terrains. Our work learns to estimate a displacement measurement from IMU data which is then fused with traditional leg odometry. Our approach greatly reduces the drift of proprioceptive state estimation, which is critical for legged robots deployed in vision and lidar denied environments such as foggy sewers or dusty mines. We compared results from an EKF and an incremental fixed-lag factor graph estimator using data from several real robot experiments crossing challenging terrains. Our results show a reduction of relative pose error by 37% in challenging scenarios when compared to a traditional kinematic-inertial estimator without learned measurement. We also demonstrate a 22% reduction in error when used with vision systems in visually degraded environments such as an underground mine