Reid Method Research Articles

Occluded Person Re-Identification Method Based on Multiscale Features and Human Feature Reconstruction

Occluded Re-ID task is proposed mainly because people are often occluded by various obstacles in the real world, which greatly affects the accuracy of model matching.In view of the challenge of occluded Re-ID, the main work of this paper is as follows :(i) Aiming at the incompleteness of human body under occlusion, an occluded Re-ID method based on multi-scale features is proposed. A partial human body locator is constructed by using the target detection algorithm to automatically recognize and cut partial human body in this method. Then this method designs a horizontal pyramid pooling strategy to extract multi-scale features and enhance the robustness of the model under the occlusion problem. Experiments show that, this method has better matching accuracy in the occluded Re-ID task. (ii) Aiming at the problem that it is difficult to align the local features between different people images under occlusion, an occluded Re-ID method based on human feature reconstruction is proposed. This method is an unaligned method, which uses sparse representation to reconstruct human body features. Difficult sample triplet loss function was improved by using human feature reconstruction distance and the proportion of similar parts to matching correlation was increased. Experiments show that this method can effectively improve the occlusion resistance of the model.

The adsorption model of the adsorption process of CH4 on coal and its thermodynamic characteristics

In this study, the adsorption and desorption data of six coal samples with different volatile fraction contents were obtained using a laboratory-made adsorption and desorption system. The experimental temperature was 30 ℃, and the adsorption pressure was 0–3.5 MPa. The fitting effects of four different calculation methods of saturated vapor pressure (P0) (namely Reid method, Antoine method, Astakhov method, and Amankwah method) in applying three adsorption models (namely two-parameter BET model, three-parameter BET model, and D-A equation) were compared and analyzed. The D-A equation was reliable as an adsorption model when the Astakhov method was used to calculate P0. When fitting the adsorption data of six coal samples, R2 values were all greater than 0.99, and the obtained fitting parameters were all within a reasonable range. The Langmuir model and D-A equation were utilized to calculate the surface potential(Ω) and Gibbs free energy change (ΔG) of each coal sample. The D-A equation was more reliable to calculate Ω and ΔG, and it had some advantages in terms of calculation accuracy.

Unsupervised domain adaptive person re-identification via camera penalty learning

Unsupervised domain adaptive (UDA) person re-identification (re-ID) aims to adapt the model trained on a labeled source domain to an unlabeled target domain. For pseudo-label-based UDA methods, pseudo label noise is the main problem for model degradation and cross-camera problem is one main factor to cause this noise. In this paper, a novel camera penalty learning (CPL) UDA person re-ID method is proposed to address this problem. The possibility of selecting wrong negative sample and positive sample is relatively high in conventional triplet loss due to cross-camera problem. To alleviate this problem, a camera-penalty-based triplet loss (PTL) is designed. It adds camera-ID-penalty to conventional triplet loss to reduce sample distance imbalance, thereby improving the quality of pseudo labels. In order to reduce the dependence on pseudo labels and improve the robustness, a camera-penalty-neighborhood loss (PNL) is designed and combined with the push loss (PL). The PNL minimizes the distance between one image and its camera-penalty-weighted neighbors. The PL maximizes the distances among all images. The proposed CPL model achieves considerable results of 87.4%/70.7% and 75.2%/59.0% Rank-1/ mAP on DukeMTMC-reID-to-Market-1501 and Market-1501-to-DukeMTMC-reID UDA tasks.

Stable Median Centre Clustering for Unsupervised Domain Adaptation Person Re-Identification.

The current unsupervised domain adaptation person re-identification (re-ID) method aims to solve the domain shift problem and applies prior knowledge learned from labelled data in the source domain to unlabelled data in the target domain for person re-ID. At present, the unsupervised domain adaptation person re-ID method based on pseudolabels has obtained state-of-the-art performance. This method obtains pseudolabels via a clustering algorithm and uses these pseudolabels to optimize a CNN model. Although it achieves optimal performance, the model cannot be further optimized due to the existence of noisy labels in the clustering process. In this paper, we propose a stable median centre clustering (SMCC) for the unsupervised domain adaptation person re-ID method. SMCC adaptively mines credible samples for optimization purposes and reduces the impact of label noise and outliers on training to improve the performance of the resulting model. In particular, we use the intracluster distance confidence measure of the sample and its K-reciprocal nearest neighbour cluster proportion in the clustering process to select credible samples and assign different weights according to the intracluster sample distance confidence of samples to measure the distances between different clusters, thereby making the clustering results more robust. The experiments show that our SMCC method can select credible and stable samples for training and improve performance of the unsupervised domain adaptation model. Our code is available at https://github.com/sunburst792/SMCC-method/tree/master.

A Novel SRTSR Model for Cross-Resolution Person Re-Identification

Compared to the general person re-identification (re-id), pedestrian images in cross-resolution person re-id tasks always have variant resolutions, thus the resolution mismatching problem between low-resolution (LR) images and high-resolution (HR) images significantly limits the performance of neural networks. The general solution is to introduce interpolation methods or super-resolution (SR) module to normalize all the images to the same resolution. Unfortunately, current SR methods can only upscale the entire image to a certain scale, but fail to unify the resolution of images with different width-to-height ratio. To solve this problem, we propose a Specific-Resolution-Targeted Super-Resolution (SRTSR) embedded person re-id method. Instead of using deep CNN-based SR methods, our SRTSR network consists of a fast and light-weight feature learning module with a self-adaptive Swish (SA-Swish) as its activation function to improve its performance, and a specific-resolution-targeted upscale module to upscale the images to any specific resolution. Compared to the traditional upscale module which can only accept the upscale factor for the entire image, our upscale module can either accept two independent and arbitrary upscale factors for either height and width, e.g., $2.4\times $ for height and $3.1\times $ for width, or any specific resolution, e.g., $256\times 128$ , helping to unify the resolution of images with different width-to-height ratio. Extensive experiments are conducted and the result shows that the performance of our method is over the SR-based state-of-the-art person re-id methods on common re-id benchmarks.

StyleGAN-LSRO Method for Person Re-Identification

In this study, the StyleGAN-LSRO method has been developed for person re-identification (re-ID) tasks. This method applies the style-based generative adversarial network (StyleGAN) to generate new synthetic images from existing person re-ID datasets and the label smoothing regularization for outliers (LSRO) algorithm to process those newly produced unlabeled images by assigning them a uniform label distribution along with the definition of a loss function for the training process. A baseline model based on a convolutional neural network (CNN) was developed to learn the discriminative features to recognize a person's identity. The developed method has been tested on three datasets. These datasets are Market-1501, DukeMTMC-reID, and MSMT17. The experimental results show that the StyleGAN model achieved a Fréchet inception distance score of 12.67 and structural similarity score of 0.387, outperforming all the previous generative methods and demonstrating that the images generated by StyleGAN are of superior quality. Adding these StyleGAN-generated data significantly improves the person re-ID accuracy. The StyleGAN-LSRO person re-ID method achieved 98.5% rank-1 accuracy and 91.8% mean average precision (mAP) on Market-1501, 87.0% rank-1 accuracy and 83.8% mAP on DukeMTMC-reID, and 81.5% rank-1 accuracy and 60.9% mAP on MSMT17, respectively. These results show that the StyleGAN-LSRO method significantly outperforms most of the state-of-the-art person re-ID methods. The success rate for person re-ID increases when the images used are of high resolution and square matrix form. In other cases, the success rate decreases.

Isobaric heat capacities of a methane (1) + propane (2) mixture by differential scanning calorimetry at near-critical and supercritical conditions

Isobaric heat capacity data are needed to test and improve thermodynamic models of natural gas over wide ranges of temperature and pressure. Measurements are reported here at temperatures between (184 and 421) K and pressures between (5 and 32) MPa for supercritical mixtures of methane (1) + propane (2) at x1 = 0.950 (±0.005). Estimated relative uncertainties in the measured heat capacities range from (1.8 to 4.5) %. In addition, measurements at temperatures of (184, 190, 197, 203 and 209) K were performed at pressures (1.67 to 2.39) MPa higher than saturation conditions to estimate the heat capacity of the mixture at the bubble point. The binary mixture data were compared with the predictions of three models: the Groupe Européen de Recherches Gazières (GERG) 2008 multi-parameter equation of state (EOS), the Peng-Robinson (PR) EOS used widely by chemical engineers, and the Statistical Associating Fluid Theory (SAFT)-γ Mie EOS incorporating group contributions. Among the three models, the PR EOS was found to describe the heat capacity values best. A brief investigation indicated that the Joback and Reid method that the SAFT-γ Mie EOS is based on fails to accurately predict the ideal gas heat capacity of methane at cryogenic temperatures, while the inaccuracy of the GERG-2008 EOS stems from the residual part of the methane (1) + propane (2) mixing functions.

Modeling the Solubility of Hydrogen Sulfide in Ionic Liquids Using van der Waals Equation of State

This study mathematically models the solubility of hydrogen sulfide in ten ionic liquids including 1-(2-hydroxyethyl)-3-imidazolium [Hoemim]+ with three different anions, 1-ethyl-3-methylimidazolium [emim]+ with two different anions, 1-hexyl-3-methylimidazolium [hmim]+ with two different anions, and 1-butyl-3-methylimidazolium [bmim]+ with three different anions. The modeling was performed by van der Waals (vdW) equation of state as a φ-φ approach in which the modified van der Waals–Berthelot mixing rule is used. The critical properties of ionic liquids have been estimated by a group of supplementary methods suggested by Valderrama and Robles, known as the modified Lyderson–Joback–Reid method. It is concluded that since the absolute average relative deviation is less than 1%, the developed model has a high accuracy.

Visual Features with Spatio-Temporal-Based Fusion Model for Cross-Dataset Vehicle Re-Identification

Vehicle re-identification (Re-Id) is the key module in an intelligent transportation system (ITS). Due to its versatile applicability in metropolitan cities, this task has received increasing attention these days. It aims to identify whether the specific vehicle has already appeared over the surveillance network or not. Mostly, the vehicle Re-Id method are evaluated on a single dataset, in which training and testing of the model is performed on the same dataset. However in practice, this negatively effects model generalization ability due to biased datasets along with the significant difference between training and testing data; hence, the model becomes weak in a practical environment. To demonstrate this issue, we have empirically shown that the current vehicle Re-Id datasets are usually strongly biased. In this regard, we also conduct an extensive study on the cross and the same dataset to examine the impact on the performance of the vehicle Re-Id system, considering existing methods. To address the problem, in this paper, we have proposed an approach with augmentation of the training dataset to reduce the influence of pose, angle, camera color response, and background information in vehicle images; whereas, spatio-temporal patterns of unlabelled target datasets are learned by transferring siamese neural network classifiers trained on a source-labelled dataset. We finally calculate the composite similarity score of spatio-temporal patterns with siamese neural-network-based classifier visual features. Extensive experiments on multiple datasets are examined and results suggest that the proposed approach has the ability to generalize adequately.

From pedestrian to group retrieval via siamese network and correlation

In many public security applications such as anomaly detection, it is important to re-identify a group of pedestrians by other surveillance cameras, which ascribes to the group retrieval problem. Most previous studies focus on single-person re-identification (re-id) and ignore the correlations among group members, and they lack a large and comprehensive group retrieval benchmark to associate these two tasks. To address this issue, this paper focuses on solving the group retrieval problem and uses it to improve re-id. First, the paper build a comprehensive benchmark for both group retrieval and the group-aided re-id task by proposing a novel pedestrian group retrieval dataset named “SYSU-Group” and a corresponding group-associated re-id dataset named “Group-reID”, which introduces realistic challenges such as variations of pose, viewpoint, illumination, and intra-group layout. The paper then proposes the Siamese Verification-Identification-based Group Retrieval (SVIGR) method, which combines verification and identification modules in a Siamese network to extract robust person features and follows the principle of minimum distance matching to realize group retrieval. Finally, a group-guided re-id method named group retrieval correlation (GRC) is proposed to improve re-id with additional group information. Experimental results on three various group retrieval benchmarks demonstrate the superiority and effectiveness of our method.

Deep manifold clustering based optimal pseudo pose representation (DMC-OPPR) for unsupervised person re-identification

Person re-identification (re-ID) is highly complex in a diverse surveillance environment. The existing person re-ID methods are evaluated as a closed set problem with limited environmental variation. It is highly challenging to estimate the diverse poses of a dynamically crowded environment using the traditional unsupervised person re-ID methods. To resolve this issue of handling complex diverse poses and camera angles, a contextual incremental multi-clustering based unsupervised person re-ID method have been proposed. Cam-pose based optimal similarity distance threshold is determined to label the unlabeled person re-ID images efficiently. Frequent intra and inter-camera pseudo pose sequences are represented with optimal distance threshold. This resolves the over-fitting issue created by the dominant samples of an identity and reduces the source-target domain gap. The experimental results show the supremacy of our proposed method over the existing unsupervised person re-ID methods in handling complex poses and camera angles in an incremental self-learning diverse surveillance environment.

Machine Estimation of Drug Melting Properties and Influence on Solubility Prediction.

There has been much recent interest in machine learning (ML) and molecular quantitative structure property relationships (QSPR). The present research evaluated modern ML-based methods implemented in commercial software (COSMOquick and Molecular Modeling Pro), compared to a classical group contribution approach (Joback and Reid method), to estimate melting points and enthalpy of fusion values. A broad data set of market compounds was gathered from the literature, together with new data measured by differential scanning calorimetry for drug candidates. The highest prediction accuracy was achieved by QSPR using stochastic gradient boosting. The model deviations were discussed, particularly the implications on thermodynamic solubility modeling, as this typically requires estimation of both melting point and enthalpy of fusion. The results suggested that despite considerable advancement in prediction accuracy, there are still limitations especially with complex drug candidates. It is recommended that in such cases, melting properties obtained in silico should be used carefully as input data for thermodynamic solubility modeling. Future research will show how the prediction limits of thermophysical drug properties can be further advanced by even larger data sets and other ML algorithms or also by using molecular simulations.

CO2 Solubility in 1,1,2,2-Tetrafluoroethanesulfonate Anion-Based Ionic Liquids: [EMIM][TFES], [BMIM][TFES], and [BNMIM][TFES

The solubility of carbon dioxide (CO2) in three different ionic liquids (ILs): 1-ethyl-3-methylimidazolium 1,1,2,2-tetrafluoroethanesulfonate ([EMIM][TFES]), 1-butyl-3-methylimidazolium 1,1,2,2-tetrafluoroethanesulfonate ([BMIM][TFES]), and 1-benzyl-3-methylimidazolium 1,1,2,2-tetrafluoroethanesulfonate ([BNMIM][TFES]) were measured. Measurements were done at temperatures ranging from 303.2 to 373.2 K (323.2 to 373.2 K for [BNMIM][TFES]) at pressures from 0.80 to 36.80 MPa. Measured data show that pressure, temperature, and characteristics of cation of ILs ([EMIM], [BMIM], and [BNMIM]) affect its CO2 solubility. We determined the CO2 solubility of fixed CO2 mole fractions for three ionic liquids. Among the three ILs, CO2 solubility intensity was ranked [BMIM][TFES] > [EMIM][TFES] ≈ [BNMIM][TFES]. The Peng–Robinson equation of state, the conventional van der Waals one-fluid mixing rule, and the modified Lydersen–Joback–Reid method were used to correlate the experimental data. The average absolute deviations of pressure were 0.0164 for CO2 + [EMIM][TFES], 0.0361 for CO2 + [BMIM][TFES], and 0.0251 for CO2 + [BNMIM][TFES] systems.

Person Search by Separated Modeling and A Mask-Guided Two-Stream CNN Model.

In this work, we tackle the problem of person search, which is a challenging task consisted of pedestrian detection and person re-identification (re-ID). Instead of sharing representations in a single joint model, we find that separating detector and re-ID feature extraction yields better performance. In order to extract more representative features for each identity, we segment out the foreground person from the original image patch. We propose a simple yet effective re-ID method, which models foreground person and original image patches individually, and obtains enriched representations from two separate CNN streams. We also propose a Confidence Weighted Stream Attention method which further re-adjusts the relative importance of the two streams by incorporating the detection confidence. Furthermore, we simplify the whole pipeline by incorporating semantic segmentation into the re-ID network, which is trained by bounding boxes as weakly-annotated masks and identification labels simultaneously. From the experiments on two standard person search benchmarks i.e. CUHK-SYSU and PRW, we achieve mAP of 83.3% and 32.8% respectively, surpassing the state of the art by a large margin. The extensive ablation study and model inspection further justifies our motivation.

Jointly Learning Commonality and Specificity Dictionaries for Person Re-Identification

Despite advances in person re-identification (re-ID), it is still far from meeting the needs of real-world applications due to tremendous visual ambiguity in person appearance across cameras. To overcome this problem, we propose a person re-ID method by decomposing a pedestrian’s appearance feature into different components. This method assumes that each pedestrian image is composed of person-shared components that reflect the similarities of different pedestrians and person-specific components that reflect unique identity information. Based on this assumption, we propose to reduce the ambiguity in visual features by removing person-shared components from pedestrian visual features. To this end, we develop a framework for learning a pair of commonality and specificity dictionaries, while introducing a distance constraint to force the particularities of the same person over the specificity dictionary to have the same coding coefficients and the coding coefficients of different pedestrians to have weak correlation. Furthermore, considering the similarity of the commonality dictionary and the sparsity of the specificity dictionary, low-rank and sparse regularization terms are introduced into the dictionary learning framework to improve their representation ability and discriminative ability. Extensive experimental results show that the proposed algorithm outperforms or is competitive with the state-of-the-art methods.

End to End Person Re-Identification Based on Attention Mechanism

This paper treats person re-identification (re-id) as a sequential model, guide person re-id with person detection, combines recurrent neural network (RNN) with attention mechanism, and proposed an end to end person re-id method for surveillance scenarios. The feature of target person is firstly extracted using ResNet, and then the feature is added to the long short-term memory (LSTM) network to guide the attention model for the region of interest in the surveillance image. Finally, combined with the information observed from the image multiple times, the most similar candidate person in the image is deduced, and the feature distance is calculated and ranked for person re-id. This paper strengthens the relationship between person detection and person re-id, and reduces the error between models. Because the number of candidate person matched with target person is reduced, this method can process person re-id task with less calculation and time. This paper also verified the effectiveness of the proposed method by experiments comparing a variety of person detection and re-id methods on several person re-id datasets.

Estimation of the acentric factor of organic compounds via a new group contribution method

To predict the acentric factor of organic compounds, a new second order group contribution method is developed. Some changes are made in the functional groups of Joback and Reid method and these groups are employed as the first order functional groups in this work. Besides, some second order functional groups are considered to improve the accuracy of estimations and distinguish among most structural isomers and stereoisomers. A set of experimental data for the acentric factor of 608 organic compounds containing complex and heavy polycyclic aromatic components are used to determine the group contribution values of 122 functional groups. The collected compounds in this research are hydrocarbons and hydrocarbon derivatives include carbon, hydrogen, oxygen, nitrogen, sulfur, fluorine, chlorine, bromine, and iodine atoms. The MATLAB software is utilized to develop the new model. The average absolute deviation of acentric factor predictions for 608 organic compounds is 0.022 and the percentage of average absolute relative deviation is 4.9%. The results of the new method are compared to the Constantinou, Gani and O'Connell method. It is observed that there is a considerable difference in the accuracy of the new method in comparison to that of the mentioned method.

Estimation of the critical constants of organic compounds via a new group contribution method

A new second order group contribution method for the estimation of critical constants of organic compounds is developed. The functional groups of Joback and Reid method are employed with some changes as first order functional groups and some second order functional groups are also defined and added to them. A set of experimental data for the critical temperature of 969 compounds, the critical pressure of 715 compounds and the critical volume of 539 compounds are collected in this study. The group contribution values of 122 functional groups have been optimized using the experimental data of organic compounds, including complex and heavy polycyclic aromatic components. The critical constants of most structural isomers and stereoisomers are distinguished in this method by considering the second order functional groups. The hydrocarbons and hydrocarbon derivatives considered in this research include carbon, hydrogen, oxygen, nitrogen, sulfur, fluorine, chlorine, bromine, and iodine atoms. Development of the new models and optimizing the values for the group contributions have been performed by the fminsearch algorithm in MATLAB software. The results of the new method are compared to three well known group contribution methods. The percentage of the average absolute relative deviations are 1.12% for critical temperature, 2.69% for critical pressure, and 2.67% for critical volume.

Person re-identification based on linear classification margin

The core of the person re-identification task is to find a discriminative feature map or a good measure to measure the similarity between two pedestrians. Under the traditional softmax loss function constraint, the characteristics of the FC layer output of the convolutional neural network are only separable and the discriminative ability is insufficient. Our proposes a person re-ID method based on decision margin. By adding the traditional softmax loss function to the angle decision margin, the characteristics of strong discriminative power in network learning are enhanced. This paper provides a clear geometric interpretation by normalizing the last fully connected layer parameters and pedestrian characteristics of the network and projecting the embedded features into the geometric space for person re-identification. Combining two different classification margins, a linear classification margin is proposed and good performance is achieved. This paper validates the effectiveness of the algorithm on the mainstream databases of Market1501 and DukeMTMC-reID. The results show that our method can get more discriminative features, and its experimental performance is significantly improved compared to the existing advanced algorithms.

Person Re-Identification Based on DropEasy Method

Currently, majority of person re-identification (reID) technologies are network-constrained by Dropout regularization, which relies on the random zeroing out of some features to make these features more independent. However, such random zeroing regularization methods are not very effective for improving network performance, because they neglect the unique contribution of different features to the network performance. To improve the value of indiscriminative features in network training, a DropEasy-based person reID method is proposed in this paper. Features are classified into discriminative and indiscriminative ones, according to the distance between the feature vectors of positive or negative sample pairs wherein the discriminative features are zeroed out, while the indiscriminative features are reserved, and the network only learns through indiscriminative features. Furthermore, because networks are always inclined to make up for incomplete information by drawing on the surrounding features in the feature maps, Dropout loses its effectiveness for the network-constraints. To solve this challenge, the DropEasy2d method that can be effectively applied to convolutional layers is further proposed in this paper. DropEasy2d searches discriminative feature areas in the feature maps by sliding and zeroing windows while reserving the indiscriminative features areas to constrain network learning. The effectiveness of the proposed method is demonstrated using the Market-1501, DuckMTMC-reID, and CUHK03 datasets. For example, in the Market-1501 dataset, DropEasy can improve the mean average precision (mAP) and Rank-1 accuracy of the ID-discriminative embedding (IDE) to 72.7%(+8.8%) and 90.5%(+6.8%), respectively, while DropEasy2d can raise them to 68.5%(+4.6%) and 88.7%(+5.0%), respectively. Based on the results, the proposed method can improve the network performance during the extraction and generalization of the discriminative features.