Georgia Gkioxari, Ross B. Girshick, Piotr Dollár
et al.
To understand the visual world, a machine must not only recognize individual object instances but also how they interact. Humans are often at the center of such interactions and detecting human-object interactions is an important practical and scientific problem. In this paper, we address the task of detecting (human, verb, object) triplets in challenging everyday photos. We propose a novel model that is driven by a human-centric approach. Our hypothesis is that the appearance of a person - their pose, clothing, action - is a powerful cue for localizing the objects they are interacting with. To exploit this cue, our model learns to predict an action-specific density over target object locations based on the appearance of a detected person. Our model also jointly learns to detect people and objects, and by fusing these predictions it efficiently infers interaction triplets in a clean, jointly trained end-to-end system we call InteractNet. We validate our approach on the recently introduced Verbs in COCO (V-COCO) and HICO-DET datasets, where we show quantitatively compelling results.
Yu-Wei Chao, Yunfan Liu, Michael Xieyang Liu
et al.
We study the problem of detecting human-object interactions (HOI) in static images, defined as predicting a human and an object bounding box with an interaction class label that connects them. HOI detection is a fundamental problem in computer vision as it provides semantic information about the interactions among the detected objects. We introduce HICO-DET, a new large benchmark for HOI detection, by augmenting the current HICO classification benchmark with instance annotations. To solve the task, we propose Human-Object Region-based Convolutional Neural Networks (HO-RCNN). At the core of our HO-RCNN is the Interaction Pattern, a novel DNN input that characterizes the spatial relations between two bounding boxes. Experiments on HICO-DET demonstrate that our HO-RCNN, by exploiting human-object spatial relations through Interaction Patterns, significantly improves the performance of HOI detection over baseline approaches.
Oriented object detection in aerial images has received extensive attention due to its wide range of application scenarios. Although great success has been achieved, current methods still suffer from inferior high-precision detection performance. Firstly, the classification scores cannot truly represent the localization accuracy of the predictions. Secondly, the orientation prediction in these detectors is not accurate enough for high-precision object detection. In this paper, we propose a T ask I nterleaving and O rientation E stimation Det ector (TIOE-Det) for high-quality oriented object detection in aerial images. Specifically, a posterior hierarchical alignment (PHA) label is proposed to optimize the detection pipeline. TIOE-Det adopts PHA label to integrate fine-grained posterior localization guidance into classification task to address the misalignment between classification and localization subtasks. Then, a balanced alignment loss is developed to solve the imbalance localization loss contribution in PHA prediction. Moreover, we propose a progressive orientation estimation (POE) strategy to approximate the orientation of objects with n-ary codes. On this basis, an angular deviation weighting strategy is proposed to achieve accurate evaluation of angle deviation in POE strategy. TIOE-Det achieves significant gains on high-precision detection performance. Extensive experiments on multiple datasets prove the superiority of our approach.
To improve the precision of defect categorization and localization in images, this paper proposes an approach for detecting surface defects in hot-rolled steel strips. The approach uses an improved YOLOv5 network model to overcome the issues of inadequate feature extraction capacity and suboptimal feature integration when identifying surface defects on steel strips. The proposed method achieves higher detection accuracy and localization precision, making it more competitive and applicable in real production. Firstly, the multi-scale feature fusion (MSF) strategy is utilized to fuse shallow and deep features effectively and enrich detailed information relevant to target defects. Secondly, the CSPLayer Res2Attention block (CRA block) residual module is introduced to reduce the loss of defect information during hierarchical transmission, thereby enhancing the extraction of fine-grained features and improving the perception of details and global features. Finally, the experimental results indicate that the mAP on the NEU-DET and GC10-DET datasets approaches 78.5% and 67.3%, respectively, which is 4.9% and 2.1% higher than that of the baseline. Meanwhile, it has higher precision and more precise localization capabilities than other methods. Furthermore, it also achieves 59.2% mAP on the APDDD dataset, indicating its potential for growth in further domains.
Among the various types of enzyme-based biosensors, sensors utilizing enzymes capable of direct electron transfer (DET) are recognized as the most ideal. However, only a limited number of redox enzymes are capable of DET with electrodes, that is, dehydrogenases harboring a subunit or domain that functions specifically to accept electrons from the redox cofactor of the catalytic site and transfer the electrons to the external electron acceptor. Such subunits or domains act as built-in mediators for electron transfer between enzymes and electrodes; consequently, such enzymes enable direct electron transfer to electrodes and are designated as DET-type enzymes. DET-type enzymes fall into several categories, including redox cofactors of catalytic reactions, built-in mediators for DET with electrodes and by their protein hierarchic structures, DET-type oxidoreductases with oligomeric structures harboring electron transfer subunits, and monomeric DET-type oxidoreductases harboring electron transfer domains. In this review, we cover the science of DET-type oxidoreductases and their biomedical applications. First, we introduce the structural biology and current understanding of DET-type enzyme reactions. Next, we describe recent technological developments based on DET-type enzymes for biomedical applications, such as biosensors and biochemical energy harvesting for self-powered medical devices. Finally, after discussing how to further engineer and create DET-type enzymes, we address the future prospects for DET-type enzymes in biomedical engineering. Expected final online publication date for the Annual Review of Biomedical Engineering, Volume 26 is May 2024. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.
This work addresses a new challenge of understanding human nonverbal interaction in social contexts. Nonverbal signals pervade virtually every communicative act. Our gestures, facial expressions, postures, gaze, even physical appearance all convey messages, without anything being said. Despite their critical role in social life, nonverbal signals receive very limited attention as compared to the linguistic counterparts, and existing solutions typically examine nonverbal cues in isolation. Our study marks the first systematic effort to enhance the interpretation of multifaceted nonverbal signals. First, we contribute a novel large-scale dataset, called NVI, which is meticulously annotated to include bounding boxes for humans and corresponding social groups, along with 22 atomic-level nonverbal behaviors under five broad interaction types. Second, we establish a new task NVI-DET for nonverbal interaction detection, which is formalized as identifying triplets in the formfrom images. Third, we propose a nonverbal interaction detection hypergraph (NVI-DEHR), a new approach that explicitly models high-order nonverbal interactions using hypergraphs. Central to the model is a dual multi-scale hypergraph that adeptly addresses individual-to-individual and group-to-group correlations across varying scales, facilitating interactional feature learning and eventually improving interaction prediction. Extensive experiments on NVI show that NVI-DEHR improves various baselines significantly in NVI-DET. It also exhibits leading performance on HOI-DET, confirming its versatility in supporting related tasks and strong generalization ability. We hope that our study will offer the community new avenues to explore nonverbal signals in more depth.
Digital extension tools (DETs) include phone calls, WhatsApp groups and specialised smartphone applications used for agricultural knowledge brokering. We researched processes through which DETs have (and have not) been used by farmers and other extension actors in low- and middle-income countries. We interviewed 40 DET developers across 21 countries and 101 DET users in Bihar, India. We found DET use is commonly constrained by fifteen pitfalls (unawareness of DET, inaccessible device, inaccessible electricity, inaccessible mobile network, insensitive to digital illiteracy, insensitive to illiteracy, unfamiliar language, slow to access, hard to interpret, unengaging, insensitive to user's knowledge, insensitive to priorities, insensitive to socio-economic constraints, irrelevant to farm, distrust). These pitfalls partially explain why women, less educated and less wealthy farmers often use DETs less, as well as why user-driven DETs (e.g. phone calls and chat apps) are often used more than externally-driven DETs (e.g. specialised smartphone apps). Our second key finding was that users often made - not just found - DETs useful for themselves and others. This suggests the word ‘appropriation’ conceptualises DET use more accurately and helpfully than the word ‘adoption’. Our final key finding was that developers and users advocated almost ubiquitously for involving desired users in DET provision. We synthesise these findings in a one-page framework to help funders and developers facilitate more useable, useful and positively impactful DETs. Overall, we conclude developers increase DET use by recognizing users as fellow developers – either through collaborative design or by designing adaptable DETs that create room for user innovation.
This paper presents a formulation alongside a numerical solution algorithm to describe the mechanical response of bodies made of a large class of viscoelastic materials undergoing arbitrary quasistatic finite deformations. With the objective of having a unified formulation that applies to a wide range of highly compressible, nearly incompressible, and fully incompressible soft organic materials in a numerically tractable manner, the viscoelasticity is described within a Lagrangian setting by a two-potential mixed formulation. In this formulation, the deformation field, a pressure field that ensues from a Legendre transform, and an internal variable of state Fv that describes the viscous part of the deformation are the independent fields. Consistent with the experimental evidence that viscous deformation is a volume-preserving process, the internal variable Fv is required to satisfy the constraint det Fv=1. To solve the resulting initial-boundary-value problem, a numerical solution algorithm is proposed that is based on a finite-element (FE) discretization of space and a finite-difference discretization of time. Specifically, a Variational Multiscale FE method is employed that allows for an arbitrary combination of shape functions for the deformation and pressure fields. To deal with the challenging non-convex constraint det Fv=1, a new time integration scheme is introduced that allows to convert any explicit or implicit scheme of choice into a stable scheme that preserves the constraint det Fv=1 identically. A series of test cases is presented that showcase the capabilities of the proposed formulation.
We consider the problem of exact integration of the TT¯\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ T\overline{T} $$\end{document} -deformation of two dimensional quantum field theories, as well as some higher dimensional extensions in the form of det T -deformations. When the action can be shown to only depend algebraically on the background metric the solution of the deformation equation on the Lagrangian can be given in closed form in terms of solutions of the (extended) Burgers’ equation. We present such examples in two and higher dimensions.
This paper reports the main physical and chemical properties of NEG materials, sorption mechanisms and use of NEG pumps from high to extreme high vacuums.
This write-up explains the signal generation mechanism in CdTe semiconductor sensors. Derivations are mostly carried out explicitly, starting with basic semiconductor relations. The analysis is largely applicable to any semiconductor, with the focus being on the Schottky-type CdTe:Cl sensors that are employed in the Spectrometer/Telescope for Imaging X-rays (STIX) instrument on-board the ESA Solar Orbiter mission.
The simulation study of the light yield and attenuation in the plastic scintillator was performed. The wavelength shifting fiber readout was embedded in the grooves machined along the entire strip surface. The scintillator strips was irradiated with a radiation source {}^{90}Sr or cosmic muons along and across the strip.
The paper reviews the characterization's techniques for solid targets used in nuclear physics with special emphasis on actinide targets. The determination of the thickness, isotopic and chemical composition are described for actinide targets during their fabrication process. Accent is given on their monitoring during irradiation.
Charlott e Appel og Nina Christensen: Avenues to Knowledge about Children’s Books and Reading 1750-1850 via The Royal Library Based on the Royal Danish Library’s collections, the article identifies paths to reading material published for children in Denmark c. 1750-1850, as well as to children’s experiences with books and the world of books in homes, schools and on the book market. This was a period of major changes with regard to the number of publications for children, to subject choices and equipment, and to the reading cultures, in which children took part. By focusing on the use of books in practice, including translated literature, books in foreign languages, as well as reading material other than books, this article presents an alternative to previous research, which prioritised specific media and genres, especially fiction and first editions of books by Danish authors. The article begins with a brief introduction to the phenomenon of doing archival research, and Maria Tamboukou’s concept of ‘researcher’s cut’ is introduced in dialogue with the term ‘research narratives’. These concepts can form the basis for reflections on how new narratives as well as new archives are created by researchers when interacting with space and matter in the archive. The first main section deals with books for children from c. 1750 to 1850 in the Royal Danish Library. ‘Childrens books’ did not exist as a separate category at the time when the systematic catalogues of the library were created, but the researchers’ establishment of a new database has made it possible to identify and present a much larger corpus of books for children than previously known. The second main section shows how it is possible to gather information about printed matter not held by the Royal Danish Library. In the third main section, it is demonstrated how a wish to identify sources to book usage in practice has led to some of the library’s special collections, including the Manuscript Collection and the Collection of Map, Pictures and Photographs, as well as to collections only recently merged with the Royal Danish Library. Autobiographies are also presented as an important source, not least when it comes to understanding the use of books by children with different social backgrounds. Finally, it is pointed out that since books for children should be perceived and studied as a transnational phenomenon, the systematic digitalisation of the publications will be of crucial importance to future Danish and international research, as well as to teaching and dissemination on the basis of this previously underexposed part of the Royal Danish Library.
Trine Brox: Text, technology, trope Tibetan Buddhist practices include the engagement with Buddhist script in the form of tightly packed scrolls of paper that are placed in drums and spun in order to accumulate merit. On the scrolls are written or printed mantras (sacred or magic sounds in the form of a series of syllables) or dhāraṇīs (formula condensing lengthy texts or teachings). The receptacles containing such scrolls are the iconic prayer wheels, whose materiality enables and restricts particular modes of textual engagement. What kind of texts are these scrolls? How does one read texts that come in the form of tightly packed scrolls? They cannot be read by immersing oneself in the content of the text because its materiality restricts access to it. Instead it fits perfectly with the prayer wheel technology that allows circumambulatory recitation, the article argues. The prayer wheel is not only an important object in recitation practices for Tibetan Buddhists, but has also become a key marker of Tibetan identity. Yet its history, culture and practice have received very little scholarly attention. This article attempts to eradicate this blind spot. It endeavors to show how the prayer wheel has earned itself its iconic status by explaining how the prayer wheel is a receptacle of sacred script and a device for reciting sacred script. The article zooms in on (i) the cult of the book in the Tibetan culture sphere, (ii) the technology of prayer wheels; and (iii) the recurring tropes of the wheel, circumambulation, and rotation in Buddhism, as well as the merit connected with them. In view of this particular constellation of book cult, technology, and dominant trope, it makes sense that, first of all, Tibetan Buddhists have adopted and further developed a technology that optimizes interaction with sacred script; secondly, that rotation is considered an adequate way to interact with Buddha’s doctrine; and third, that this device has become an icon for the Tibetan civilization. It is especially the sacred text within the wheel, the article argues, that endows the prayer wheel with high status in a hierarchy of Buddhist material objects. Finally, this raises questions about modernized prayer wheels – when technological progress has enabled further development of devices that can contain and spin Buddhist script such as optic discs, prayer wheel apps, and automated praying machines. How does their materiality impact textual engagement? The article is based upon data produced through (i) ethnographic inquiry, such as interacting with the stakeholders who deal with Buddhist material culture for different reasons, e.g. producers, marketers, ritual specialists, practitioners, and consumers, and (ii) textual sources that includes Tibetan and English-language scholarship, catalogues, user’s guides, and marketing material. This multi-modal method has produced knowledge about the prayer wheel as practice, i.e. what we can call popular religion, and as theory, i.e. according to how the prayer wheel is idealized in the writings of Buddhist masters.
In this paper, we review and extend a family of log-det divergences for symmetric positive definite (SPD) matrices and discuss their fundamental properties. We show how to generate from parameterized Alpha-Beta (AB) and Gamma Log-det divergences many well known divergences, for example, the Stein's loss, S-divergence, called also Jensen-Bregman LogDet (JBLD) divergence, the Logdet Zero (Bhattacharryya) divergence, Affine Invariant Riemannian Metric (AIRM) as well as some new divergences. Moreover, we establish links and correspondences among many log-det divergences and display them on alpha-beta plain for various set of parameters. Furthermore, this paper bridges these divergences and shows also their links to divergences of multivariate and multiway Gaussian distributions. Closed form formulas are derived for gamma divergences of two multivariate Gaussian densities including as special cases the Kullback-Leibler, Bhattacharryya, R\'enyi and Cauchy-Schwartz divergences. Symmetrized versions of the log-det divergences are also discussed and reviewed. A class of divergences is extended to multiway divergences for separable covariance (precision) matrices.