scholarly journals Instance Hard Triplet Loss for In-video Person Re-identification

2020 ◽  
Vol 10 (6) ◽  
pp. 2198
Author(s):  
Xing Fan ◽  
Wei Jiang ◽  
Hao Luo ◽  
Weijie Mao ◽  
Hongyan Yu
Keyword(s):  
Loss Function ◽  
Large Scale ◽  
Video Camera ◽  
Human Tracking ◽  
Training Time ◽  
Pose Tracking ◽  
Person Features ◽  
Potential Applications ◽  
Triplet Loss ◽  
Video Group

Traditional Person Re-identification (ReID) methods mainly focus on cross-camera scenarios, while identifying a person in the same video/camera from adjacent subsequent frames is also an important question, for example, in human tracking and pose tracking. We try to address this unexplored in-video ReID problem with a new large-scale video-based ReID dataset called PoseTrack-ReID with full images available and a new network structure called ReID-Head, which can extract multi-person features efficiently in real time and can be integrated with both one-stage and two-stage human or pose detectors. A new loss function is also required to solve this new in-video problem. Hence, a triplet-based loss function with an online hard example mining designed to distinguish persons in the same video/group is proposed, called instance hard triplet loss, which can be applied in both cross-camera ReID and in-video ReID. Compared with the widely-used batch hard triplet loss, our proposed loss achieves competitive performance and saves more than 30% of the training time. We also propose an automatic reciprocal identity association method, so we can train our model in an unsupervised way, which further extends the potential applications of in-video ReID. The PoseTrack-ReID dataset and code will be publicly released.

scholarly journals A Feature Extraction Method for Person Re-identification Based on a Two-branch CNN

2020 ◽  
Author(s):  
Linlin Li ◽  
Bo Yang ◽  
Shaohui Chen
Keyword(s):  
Feature Extraction ◽  
Loss Function ◽  
Training Process ◽  
Feature Extraction Method ◽  
Network Training ◽  
Person Features ◽  
Triplet Loss ◽  
Improved Learning ◽  
Open Datasets ◽  
Global And Local

Abstract A two-branch convolutional neural network (CNN) architecture for feature extraction in person re-identification (re-ID) based on video surveillance is proposed. Highly discriminative person features are obtained by extracting both global and local features. Moreover, an adaptive triplet loss function based on the original triplet loss function is proposed and is used in the network training process, resulting in a significantly improved learning efficiency. The experimental results on open datasets demonstrate the effectiveness of the proposed method.

scholarly journals C4 Bacterial Volatiles Improve Plant Health

Pathogens ◽  
2021 ◽  
Vol 10 (6) ◽  
pp. 682
Author(s):  
Bruno Henrique Silva Dias ◽  
Sung-Hee Jung ◽  
Juliana Velasco de Castro Oliveira ◽  
Choong-Min Ryu
Keyword(s):  
Plant Growth ◽  
Volatile Compounds ◽  
Large Scale ◽  
Growth Promotion ◽  
Plant Growth Promoting Rhizobacteria ◽  
Plant Health ◽  
Systemic Resistance ◽  
Potential Applications ◽  
Plant Growth Promoting ◽  
Bacterial Volatiles

Plant growth-promoting rhizobacteria (PGPR) associated with plant roots can trigger plant growth promotion and induced systemic resistance. Several bacterial determinants including cell-wall components and secreted compounds have been identified to date. Here, we review a group of low-molecular-weight volatile compounds released by PGPR, which improve plant health, mostly by protecting plants against pathogen attack under greenhouse and field conditions. We particularly focus on C4 bacterial volatile compounds (BVCs), such as 2,3-butanediol and acetoin, which have been shown to activate the plant immune response and to promote plant growth at the molecular level as well as in large-scale field applications. We also disc/ uss the potential applications, metabolic engineering, and large-scale fermentation of C4 BVCs. The C4 bacterial volatiles act as airborne signals and therefore represent a new type of biocontrol agent. Further advances in the encapsulation procedure, together with the development of standards and guidelines, will promote the application of C4 volatiles in the field.

scholarly journals Shape My Face: Registering 3D Face Scans by Surface-to-Surface Translation

2021 ◽  
Author(s):  
Mehdi Bahri ◽  
Eimear O’ Sullivan ◽  
Shunwang Gong ◽  
Feng Liu ◽  
Xiaoming Liu ◽  
...  
Keyword(s):  
Large Scale ◽  
Training Time ◽  
3D Face ◽  
In The Wild ◽  
Previous State ◽  
Surface Translation ◽  
Visual Attention Mechanism ◽  
Diverse Data ◽  
Human Faces ◽  
Robust To Noise

AbstractStandard registration algorithms need to be independently applied to each surface to register, following careful pre-processing and hand-tuning. Recently, learning-based approaches have emerged that reduce the registration of new scans to running inference with a previously-trained model. The potential benefits are multifold: inference is typically orders of magnitude faster than solving a new instance of a difficult optimization problem, deep learning models can be made robust to noise and corruption, and the trained model may be re-used for other tasks, e.g. through transfer learning. In this paper, we cast the registration task as a surface-to-surface translation problem, and design a model to reliably capture the latent geometric information directly from raw 3D face scans. We introduce Shape-My-Face (SMF), a powerful encoder-decoder architecture based on an improved point cloud encoder, a novel visual attention mechanism, graph convolutional decoders with skip connections, and a specialized mouth model that we smoothly integrate with the mesh convolutions. Compared to the previous state-of-the-art learning algorithms for non-rigid registration of face scans, SMF only requires the raw data to be rigidly aligned (with scaling) with a pre-defined face template. Additionally, our model provides topologically-sound meshes with minimal supervision, offers faster training time, has orders of magnitude fewer trainable parameters, is more robust to noise, and can generalize to previously unseen datasets. We extensively evaluate the quality of our registrations on diverse data. We demonstrate the robustness and generalizability of our model with in-the-wild face scans across different modalities, sensor types, and resolutions. Finally, we show that, by learning to register scans, SMF produces a hybrid linear and non-linear morphable model. Manipulation of the latent space of SMF allows for shape generation, and morphing applications such as expression transfer in-the-wild. We train SMF on a dataset of human faces comprising 9 large-scale databases on commodity hardware.

scholarly journals Review of Cryogenic Carbon Capture Innovations and Their Potential Applications

2021 ◽  
Vol 7 (3) ◽  
pp. 58
Author(s):  
Carolina Font-Palma ◽  
David Cann ◽  
Chinonyelum Udemu
Keyword(s):  
Carbon Dioxide Emissions ◽  
Carbon Capture ◽  
Large Scale ◽  
Carbon Capture And Storage ◽  
Technological Readiness ◽  
Cryogenic Carbon Capture ◽  
Potential Applications ◽  
Air Capture ◽  
Reduce Carbon Dioxide ◽  
Co2 Recovery

Our ever-increasing interest in economic growth is leading the way to the decline of natural resources, the detriment of air quality, and is fostering climate change. One potential solution to reduce carbon dioxide emissions from industrial emitters is the exploitation of carbon capture and storage (CCS). Among the various CO2 separation technologies, cryogenic carbon capture (CCC) could emerge by offering high CO2 recovery rates and purity levels. This review covers the different CCC methods that are being developed, their benefits, and the current challenges deterring their commercialisation. It also offers an appraisal for selected feasible small- and large-scale CCC applications, including blue hydrogen production and direct air capture. This work considers their technological readiness for CCC deployment and acknowledges competing technologies and ends by providing some insights into future directions related to the R&D for CCC systems.

scholarly journals Quantifying the effects of passenger-level heterogeneity on transit journey times

2020 ◽  
Vol 1 ◽  
Author(s):  
Ramandeep Singh ◽  
Daniel J. Graham ◽  
Richard J. Anderson
Keyword(s):  
Large Scale ◽  
Semiparametric Regression ◽  
Related Factors ◽  
Network Characteristics ◽  
Journey Time ◽  
Time Variance ◽  
Assignment Algorithm ◽  
Rail Systems ◽  
Urban Rail ◽  
Potential Applications

Abstract In this paper, we apply flexible data-driven analysis methods on large-scale mass transit data to identify areas for improvement in the engineering and operation of urban rail systems. Specifically, we use data from automated fare collection (AFC) and automated vehicle location (AVL) systems to obtain a more precise characterisation of the drivers of journey time variance on the London Underground, and thus an improved understanding of delay. Total journey times are decomposed via a probabilistic assignment algorithm, and semiparametric regression is undertaken to disentangle the effects of passenger-specific travel characteristics from network-related factors. For total journey times, we find that network characteristics, primarily train speeds and headways, represent the majority of journey time variance. However, within the typically twice as onerous access and egress time components, passenger-level heterogeneity is more influential. On average, we find that intra-passenger heterogeneity represents 6% and 19% of variance in access and egress times, respectively, and that inter-passenger effects have a similar or greater degree of influence than static network characteristics. The analysis shows that while network-specific characteristics are the primary drivers of journey time variance in absolute terms, a nontrivial proportion of passenger-perceived variance would be influenced by passenger-specific characteristics. The findings have potential applications related to improving the understanding of passenger movements within stations, for example, the analysis can be used to assess the relative way-finding complexity of stations, which can in turn guide transit operators in the targeting of potential interventions.

scholarly journals Deep Learning of Appearance Affinity for Multi-Object Tracking and Re-Identification: A Comparative View

Electronics ◽  
2020 ◽  
Vol 9 (11) ◽  
pp. 1757
Author(s):  
María J. Gómez-Silva ◽  
Arturo de la Escalera ◽  
José M. Armingol
Keyword(s):  
Deep Learning ◽  
Object Tracking ◽  
Loss Function ◽  
Neural Model ◽  
Training Data ◽  
Learning Approaches ◽  
The Core ◽  
Triplet Loss ◽  
Affinity Model

Recognizing the identity of a query individual in a surveillance sequence is the core of Multi-Object Tracking (MOT) and Re-Identification (Re-Id) algorithms. Both tasks can be addressed by measuring the appearance affinity between people observations with a deep neural model. Nevertheless, the differences in their specifications and, consequently, in the characteristics and constraints of the available training data for each one of these tasks, arise from the necessity of employing different learning approaches to attain each one of them. This article offers a comparative view of the Double-Margin-Contrastive and the Triplet loss function, and analyzes the benefits and drawbacks of applying each one of them to learn an Appearance Affinity model for Tracking and Re-Identification. A batch of experiments have been conducted, and their results support the hypothesis concluded from the presented study: Triplet loss function is more effective than the Contrastive one when an Re-Id model is learnt, and, conversely, in the MOT domain, the Contrastive loss can better discriminate between pairs of images rendering the same person or not.

A review of passive self-sensing tag

Sensor Review ◽  
2017 ◽  
Vol 37 (3) ◽  
pp. 338-345 ◽  
Author(s):  
Yawei Xu ◽  
Lihong Dong ◽  
Haidou Wang ◽  
Jiannong Jing ◽  
Yongxiang Lu
Keyword(s):  
Radio Frequency Identification ◽  
Large Scale ◽  
Low Cost ◽  
Integrated Sensor ◽  
Content Type ◽  
Passive Sensing ◽  
Static Information ◽  
Potential Applications ◽  
Frequency Identification ◽  
Identification Tags

Purpose Radio frequency identification tags for passive sensing have attracted wide attention in the area of Internet of Things (IoT). Among them, some tags can sense the property change of objects without an integrated sensor, which is a new trend of passive sensing based on tag. The purpose of this paper is to review recent research on passive self-sensing tags (PSSTs). Design/methodology/approach The PSSTs reported in the past decade are classified in terms of sensing mode, composition and the ways of power supply. This paper presents operation principles of PSSTs and analyzes the characteristics of them. Moreover, the paper focuses on summarizing the latest sensing parameters of PSSTs and their matching equipment. Finally, some potential applications and challenges faced by this emerging technique are discussed. Findings PSST is suitable for long-term and large-scale monitoring compared to conventional sensors because it gets rid of the limitation of battery and has relatively low cost. Also, the static information of objects stored in different PSSTs can be identified by a single reader without touch. Originality/value This paper provides a detailed and timely review of the rapidly growing research in PSST.

scholarly journals Foaming of rhamnolipids fermentation: impact factors and fermentation strategies

2021 ◽  
Vol 20 (1) ◽  
Author(s):  
Zhijin Gong ◽  
Ge Yang ◽  
Chengchuan Che ◽  
Jinfeng Liu ◽  
Meiru Si ◽  
...  
Keyword(s):  
Production Process ◽  
Large Scale ◽  
Commercial Production ◽  
Impact Factors ◽  
Scale Production ◽  
Large Scale Production ◽  
Foaming Behavior ◽  
Production Strategies ◽  
Potential Applications ◽  
And Control

AbstractRhamnolipids have recently attracted considerable attentions because of their excellent biosurfactant performance and potential applications in agriculture, environment, biomedicine, etc., but severe foaming causes the high cost of production, restraining their commercial production and applications. To reduce or eliminate the foaming, numerous explorations have been focused on foaming factors and fermentation strategies, but a systematic summary and discussion are still lacking. Additionally, although these studies have not broken through the bottleneck of foaming, they are conducive to understanding the foaming mechanism and developing more effective rhamnolipids production strategies. Therefore, this review focuses on the effects of fermentation components and control conditions on foaming behavior and fermentation strategies responded to the severe foaming in rhamnolipids fermentation and systematically summarizes 6 impact factors and 9 fermentation strategies. Furthermore, the potentialities of 9 fermentation strategies for large-scale production are discussed and some further strategies are suggested. We hope this review can further facilitate the understanding of foaming factors and fermentation strategies as well as conducive to developing the more effective large-scale production strategies to accelerate the commercial production process of rhamnolipids.

Sign in / Sign up

Export Citation Format

Share Document