Counting Datasets Research Articles

DTCC: Multi-level dilated convolution with transformer for weakly-supervised crowd counting

Crowd counting provides an important foundation for public security and urban management. Due to the existence of small targets and large density variations in crowd images, crowd counting is a challenging task. Mainstream methods usually apply convolution neural networks (CNNs) to regress a density map, which requires annotations of individual persons and counts. Weakly-supervised methods can avoid detailed labeling and only require counts as annotations of images, but existing methods fail to achieve satisfactory performance because a global perspective field and multi-level information are usually ignored. We propose a weakly-supervised method, DTCC, which effectively combines multi-level dilated convolution and transformer methods to realize end-to-end crowd counting. Its main components include a recursive swin transformer and a multi-level dilated convolution regression head. The recursive swin transformer combines a pyramid visual transformer with a fine-tuned recursive pyramid structure to capture deep multi-level crowd features, including global features. The multi-level dilated convolution regression head includes multi-level dilated convolution and a linear regression head for the feature extraction module. This module can capture both low- and high-level features simultaneously to enhance the receptive field. In addition, two regression head fusion mechanisms realize dynamic and mean fusion counting. Experiments on four well-known benchmark crowd counting datasets (UCF_CC_50, ShanghaiTech, UCF_QNRF, and JHU-Crowd++) show that DTCC achieves results superior to other weakly-supervised methods and comparable to fully-supervised methods.

The balanced discrete Burr–Hatke model and mixing INAR(1) process: properties, estimation, forecasting and COVID-19 applications

The main concern of this paper is providing a flexible discrete model that captures every kind of dispersion (equi-, over- and under-dispersion). Based on the balanced discretization method, a new discrete version of Burr–Hatke distribution is introduced with the partial moment-preserving property. Some statistical properties of the new distribution are introduced, and the applicability of proposed model is evaluated by considering counting series. A new integer-valued autoregressive (INAR) process based on the mixing Pegram and binomial thinning operators with discrete Burr–Hatke innovations is introduced, which can model contagious data properly. The different estimation approaches of parameters of the new process are provided and compared through the Monte Carlo simulation scheme. The performance of the proposed process is evaluated by four data sets of the daily death counts of the COVID-19 in Austria, Switzerland, Nigeria and Slovenia in comparison with some competitor INAR(1) models, along with the Pearson residual analysis of the assessing model. The goodness of fit measures affirm the adequacy of the proposed process in modeling all COVID-19 data sets. The fundamental prediction procedures are considered for new process by classic, modified Sieve bootstrap and Bayesian forecasting methods for all COVID-19 data sets, which is concluded that the Bayesian forecasting approach provides more reliable results.

High‐precision skeleton‐based human repetitive action counting

AbstractA novel counting model is presented by the authors to estimate the number of repetitive actions in temporal 3D skeleton data. As per the authors’ knowledge, this is the first work of this kind using skeleton data for high‐precision repetitive action counting. Different from existing works on RGB video data, the authors’ model follows a bottom‐up pipeline to clip the sub‐action first followed by robust aggregation in inference. First, novel counting loss functions and robust inference with backtracking is proposed to pursue precise per‐frame count as well as overall count with boundary frames. Second, an efficient synthetic approach is proposed to augment skeleton data in training and thus avoid time‐consuming repetitive action data collection work. Finally, a challenging human repetitive action counting dataset named VSRep is collected with various types of action to evaluate the proposed model. Experiments demonstrate that the proposed counting model outperforms existing video‐based methods by a large margin in terms of accuracy in real‐time inference.

On the performance of some new ridge parameter estimators in the Poisson-inverse Gaussian ridge regression

The Poisson Inverse Gaussian Regression model (PIGRM) is used for modeling the count datasets to deal with the issue of over-dispersion. Generally, the maximum likelihood estimator (MLE) is used to estimate the PIGRM estimates. In the PIGRM, when the explanatory variables are correlated, the MLE does not provide efficient results. To overcome this problem, we propose a ridge estimator for the PIGRM. The matrix mean square error (MSE) and the scalar MSE properties are derived and then compared with the MLE. In the ridge estimator, ridge parameter play a significant role, so, this study also proposes different ridge parameter estimators for the PIGRM. The performance of the proposed estimator is evaluated with the help of a simulation study and a real-life application using MSE as a performance evaluation criterion. The simulation study and the real-life application results show the superiority of the proposed parameter estimators as compared to the MLE.

Counting Varying Density Crowds Through Density Guided Adaptive Selection CNN and Transformer Estimation

In real-world crowd counting applications, the crowd densities in an image vary greatly. When facing density variation, humans tend to locate and count the targets in low-density regions, and reason the number in high-density regions. We observe that CNN focus on the local information correlation using a fixed-size convolution kernel and the Transformer could effectively extract the semantic crowd information by using the global self-attention mechanism. Thus, CNN could locate and estimate crowds accurately in low-density regions, while it is hard to properly perceive the densities in high-density regions. On the contrary, Transformer has a high reliability in high-density regions, but fails to locate the targets in sparse regions. Neither CNN nor Transformer can well deal with this kind of density variation. To address this problem, we propose a CNN and Transformer Adaptive Selection Network (CTASNet) which can adaptively select the appropriate counting branch for different density regions. Firstly, CTASNet generates the prediction results of CNN and Transformer. Then, considering that CNN/Transformer is appropriate for low/high-density regions, a density guided adaptive selection module is designed to automatically combine the predictions of CNN and Transformer. Moreover, to reduce the influences of annotation noise, we introduce a Correntropy based optimal transport loss. Extensive experiments on four challenging crowd counting datasets have validated the proposed method.

A Novel Discrete Generator with Modeling Engineering, Agricultural and Medical Count and Zero-Inflated Real Data with Bayesian, and Non-Bayesian Inference

This study introduces a unique flexible family of discrete probability distributions for modeling extreme count and zero-inflated count data with different failure rates. Certain significant mathematical properties, such as the cumulant generating function, moment generating function, dispersion index, L-moments, ordinary moments, and central moment are derived. The new failure rate function offers a wide range of flexibility, including “upside down”, “monotonically decreasing”, “bathtub”, “monotonically increasing” and “decreasing-constant failure rate” and “constant”. Moreover, the new probability mass function accommodates many useful shapes including the “right skewed function with no peak”, “symmetric”, “right skewed with one peak” and “left skewed with one peak”. To obtain significant characterization findings, the hazard function and the conditional expectation of certain function of the random variable are both employed. Both Bayesian and non-Bayesian estimate methodologies are considered when estimating, assessing, and comparing inferential efficacy. The Bayesian estimation approach for the squared error loss function is suggested, and it is explained. Markov chain Monte Carlo simulation studies are performed using the Metropolis Hastings algorithm and the Gibbs sampler to compare non-Bayesian vs. Bayesian results. Four real-world applications of count data sets are used to evaluate the Bayesian versus non-Bayesian techniques. Four more real count data applications are used to illustrate the significance and versatility of the new discrete class.

A lightweight network for portable fry counting devices

Estimating the number of fries plays a critical role in the maintenance of fish breeding, transportation, and the preservation of marine resources in aquaculture. Generally speaking, statistics are recorded manually by fishers and government units. Manual recording is time-consuming and increases the workload of fishers. Compared with traditional physical shunt devices, visual-based algorithms have benefits such as non-restriction of labors, minimal equipment installation, and maintenance costs. However, these methods generally come with massive calculations and model parameters, or poor abilities of aggregation handles and counting precision. This paper proposes a fry counting method named MSENet for portable fry counting devices. Firstly, the lightweight network is designed with simpler parameters (Params: 139.46 kB) for portable embedding. The visualized single-channel fry density maps are predicted by feeding the original images and the number of fries is calculated through integration. Then, the Squeeze-and-Excitation block is utilized to strengthen the features of weighty channels. The model training is refined by hyperparameter studies, the shortened preparation stage enhances the portability. What is more, a fry counting dataset NCAUF and an extra set NCAUF-EX are built for verifications of network generalization. The results demonstrate that the lightweight MSENet outperforms in fry counting with higher precision and competently solves the issue of fry aggregation (MAE: 3.33). The source code and pre-trained models are available at: https://github.com/vranlee/MSENet.

A new one-parameter discrete probability distribution with its neutrosophic extension: mathematical properties and applications.

Count data modeling's significance and its applicability to real-world occurrences have been emphasized in a number of research studies. The purpose of this work is to introduce a new one-parameter discrete distribution for the modeling of count datasets. Some mathematical properties, such as reliability measures, characteristic function, moment-generating function, and associated measurements, such as mean, variance, skewness, kurtosis, and index of dispersion, have been derived and studied. The nature of the probability mass function and failure rate function has been studied graphically. The model parameter is estimated using renowned maximum likelihood estimation methods. A neutrosophic extension of the new model is also introduced for the modeling of interval datasets. In addition, the proposed distribution's applicability was compared to that of other discrete distributions. The study's findings show that the novel discrete distribution is a very appealing alternative to some other discrete competitive distributions.

Probabilistic canonical correlation analysis for sparse count data

Canonical correlation analysis (CCA) is a classical and important multivariate technique for exploring the relationship between two sets of continuous variables. CCA has applications in many fields, such as genomics and neuroimaging. It can extract meaningful features as well as use these features for subsequent analysis. Although some sparse CCA methods have been developed to deal with high-dimensional problems, they are designed specifically for continuous data and do not consider the integer-valued data from next-generation sequencing platforms that exhibit very low counts for some important features. We propose a model-based probabilistic approach for correlation and canonical correlation estimation for two sparse count data sets. Probabilistic sparse CCA (PSCCA) demonstrates that correlations and canonical correlations estimated at the natural parameter level are more appropriate than traditional estimation methods applied to the raw data. We demonstrate through simulation studies that PSCCA outperforms other standard correlation approaches and sparse CCA approaches in estimating the true correlations and canonical correlations at the natural parameter level. We further apply the PSCCA method to study the association of miRNA and mRNA expression data sets from a squamous cell lung cancer study, finding that PSCCA can uncover a large number of strongly correlated pairs than standard correlation and other sparse CCA approaches.
 Journal of Statistical Research 2022, Vol. 56, No. 1, pp. 73-98

An Intelligent Approach for Blood Cell Detection Employing Faster RCNN

A CBC (complete blood count) is an essential component of a thorough medical evaluation. The practice of medicine has had a significant impact on popular methodologies, such as manual counting and automated analyzers. The shape of blood cells and other cellular characteristics are highly sensitive to contamination levels in the human body. Tiny blood cell images are studied to spot disease and deviations from the norm that may indicate internal contamination. Correct cell segmentation enables more precise and powerful disease detection. Examining blood cells under a microscope is an important part of any pathological investigation. It emphasizes the examination of the correct malady after pinpointing its precise location and then ranking its anomalies, which plays a crucial role in determining the nature of a patient's condition, planning treatment, and evaluating the outcome of that treatment. Initially the complete blood count (CBC) test of patient is carried out if the report suggested that there is abnormality in the blood than that blood is poured on strip and blood is stained out through addition of coagulant. Expert hematologists are in short supply, particularly in underdeveloped nations, and are frequently overwhelmed. To help them with their burden, we provide a unique method for the automated assessment of family disease using artificial intelligence on blood smear images in this paper. For object detection to be useful in real life, it needs to be quick and good at finding different things in an image. There have been many improvements made to object detection, such as Convolutional Neural Network and the family of R-CNN. In our research, we employ Faster R-CNN to look for RBCs, WBCs, and Platelets in the Blood Cell Count (CBC) and Detection dataset. The number, shape, and other meta-information, or any abnormalities in the different parts of blood, could help find problems and/or diseases like leukemia, anemia, lymphoma, sickle cell disease, thrombocytopenia, leukopenia etc., early on. Our model is accurate enough to find a bounding box on the blood parts. A box is drawn along with their name around the various components of the blood smear slide.

Spatial-Temporal Graph Network for Video Crowd Counting

In recent years, researchers have developed many deep-learning-based methods to count crowd numbers in static images. However, much fewer works focus on video-based crowd counting, in which the critical challenge of temporal correlation has not been well explored. This paper proposes a Spatial-Temporal Graph Network (STGN) to achieve efficient and accurate crowd counting in videos via learning pixel-wise and patch-wise relations in local spatial-temporal domains. Specifically, we design a pyramid graph module to leverage multi-scale features. In each scale, we sequentially construct three graphs: spatial-temporal pixel graph, temporal patch graph, and spatial pixel graph, in which we apply the self-attention mechanism to capture pixel-wise relation, learn structure-aware relation, and aggregate local features, respectively. Furthermore, we propose spatial-aware channel-wise attention to effectively fuse multi-scale features. To demonstrate the effectiveness of the proposed method, we conduct experiments on five crowd counting datasets, including a large-scale video crowd dataset (FDST). Moreover, the proposed model is also applied in the vehicle counting dataset (TRANCOS). The results show that the proposed model outperforms existing spatial-temporal crowd counting models and achieves state-of-the-art. The code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/wuzhe71/STGN</uri>

White Blood Cell Classification based on Gray Level Co-occurrence Matrix with Zero Phase Component Analysis Approach

This paper presents an approach to WBC classification by employing Gray Level Co-occurrence Matrix (GLCM) along with the Analysis of Variance (ANOVA) test and Zero Phase Component Analysis (ZCA) whitening. Moreover, the performance is evaluated through the K- Nearest Neighbor (K-NN) classifier. The proposed approach has achieved an accuracy of 94.25% on the Blood Cell Count and Detection (BCCD) dataset for the classification of four categories of WBC namely, lymphocytes, monocytes, neutrophils, and eosinophils. The experimental results reveal an improvement in accuracy (11.05, 8.15, and 14.25%) in comparison to the state-of-the-art approaches i.e., Watershed segmentation, Local Binary Pattern (LBP), and Extreme Learning Machines (ELMs) respectively.

Saddlepoint p-values for a class of nonparametric tests for the current status and panel count data under generalized permuted block design

&lt;abstract&gt;&lt;p&gt;Current status and panel count data appear in many applied fields, including medicine, clinical trials, epidemiology, econometrics, demography, engineering and public health. Therefore, in this article, we use the saddlepoint approximation method to approximate the exact p-value of a number of nonparametric tests for the current status and panel count data under a generalized permuted block design. The saddlepoint approximation is referred to as higher-order approximation and it is more accurate than the methods that lead to approximations that are accurate to the first order, such as the asymptotic normal approximation method. To verify the accuracy and efficiency of the saddlepoint approximation method, a simulation study is conducted. The simulation study results confirm that the saddlepoint approximation method is more powerful than the existing approximation method. Furthermore, number of real current status and panel count data sets are analyzed and displayed as illustrative examples.&lt;/p&gt;&lt;/abstract&gt;

Rice Plant Counting, Locating, and Sizing Method Based on High-Throughput UAV RGB Images.

Rice plant counting is crucial for many applications in rice production, such as yield estimation, growth diagnosis, disaster loss assessment, etc. Currently, rice counting still heavily relies on tedious and time-consuming manual operation. To alleviate the workload of rice counting, we employed an UAV (unmanned aerial vehicle) to collect the RGB images of the paddy field. Then, we proposed a new rice plant counting, locating, and sizing method (RiceNet), which consists of one feature extractor frontend and 3 feature decoder modules, namely, density map estimator, plant location detector, and plant size estimator. In RiceNet, rice plant attention mechanism and positive-negative loss are designed to improve the ability to distinguish plants from background and the quality of the estimated density maps. To verify the validity of our method, we propose a new UAV-based rice counting dataset, which contains 355 images and 257,793 manual labeled points. Experiment results show that the mean absolute error and root mean square error of the proposed RiceNet are 8.6 and 11.2, respectively. Moreover, we validated the performance of our method with two other popular crop datasets. On these three datasets, our method significantly outperforms state-of-the-art methods. Results suggest that RiceNet can accurately and efficiently estimate the number of rice plants and replace the traditional manual method.

Sources of variation in estimating breeding success of migratory birds from autumn counts

Abstract Understanding drivers of change in population sizes requires estimation of demographic rates such as survival and productivity. In migratory geese, productivity or breeding success is typically assessed at the autumn staging and wintering grounds by observing the number of young versus adults in flocks of geese—also called age counts. Such age counts are, however, likely to be affected by a number of factors as we are compelled to sample from an open population, in which the temporal and spatial age composition can vary due to differential migration, mortality and flocking behaviour. In this study we seek to provide guidance for the design of age counts, by identifying which factors need to be taken into account when collecting data. Identification of these factors will facilitate a more targeted data collection and enable better conservation and management recommendations. We use the long‐term age count dataset for the Svalbard population of the pink‐footed goose and focus on the following factors: May thaw days on Svalbard, region of sampling, flock size, time and cumulative harvest, calculated as the amount of shot individuals up to each observation of juveniles. We find that the temporal trend in goose productivity based on raw data will be affected by variation in how (which flock sizes), where (which region) and when (which Julian day) the data have been collected. The sources of variation in observations of goose productivity confound inference at the population level. Thus, if the goal is to follow changes in annual productivity, pooled totals of age counts are only useful if data are collected following a random design. We recommend adjusting for these effects using a demographic population model, particularly for hunted species where age counts are conducted during the hunting season and where annual population assessment work is used to recommend an optimal harvest strategy and allocation of hunting quotas. These considerations not only apply to geese, but more generally to migratory birds where breeding and non‐breeding segments or age groups partially segregate in time and space in autumn.

Likelihood ratios for categorical count data with applications in digital forensics

Abstract We consider the forensic context in which the goal is to assess whether two sets of observed data came from the same source or from different sources. In particular, we focus on the situation in which the evidence consists of two sets of categorical count data: a set of event counts from an unknown source tied to a crime and a set of event counts generated by a known source. Using a same-source versus different-source hypothesis framework, we develop an approach to calculating a likelihood ratio. Under our proposed model, the likelihood ratio can be calculated in closed form, and we use this to theoretically analyse how the likelihood ratio is affected by how much data is observed, the number of event types being considered, and the prior used in the Bayesian model. Our work is motivated in particular by user-generated event data in digital forensics, a context in which relatively few statistical methodologies have yet been developed to support quantitative analysis of event data after it is extracted from a device. We evaluate our proposed method through experiments using three real-world event datasets, representing a variety of event types that may arise in digital forensics. The results of the theoretical analyses and experiments with real-world datasets demonstrate that while this model is a useful starting point for the statistical forensic analysis of user-generated event data, more work is needed before it can be applied for practical use.

A Phylogenetic Framework to Simulate Synthetic Interspecies RNA-Seq Data.

Interspecies RNA-Seq datasets are increasingly common, and have the potential to answer new questions about the evolution of gene expression. Single-species differential expression analysis is now a well-studied problem that benefits from sound statistical methods. Extensive reviews on biological or synthetic datasets have provided the community with a clear picture on the relative performances of the available methods in various settings. However, synthetic dataset simulation tools are still missing in the interspecies gene expression context. In this work, we develop and implement a new simulation framework. This tool builds on both the RNA-Seq and the phylogenetic comparative methods literatures to generate realistic count datasets, while taking into account the phylogenetic relationships between the samples. We illustrate the usefulness of this new framework through a targeted simulation study, that reproduces the features of a recently published dataset, containing gene expression data in adult eye tissue across blind and sighted freshwater crayfish species. Using our simulated datasets, we perform a fair comparison of several approaches used for differential expression analysis. This benchmark reveals some of the strengths and weaknesses of both the classical and phylogenetic approaches for interspecies differential expression analysis, and allows for a reanalysis of the crayfish dataset. The tool has been integrated in the R package compcodeR, freely available on Bioconductor.

Locating and Counting Heads in Crowds With a Depth Prior.

To simultaneously estimate the number of heads and locate heads with bounding boxes, we resort to detection-based crowd counting by leveraging RGB-D data and design a dual-path guided detection network (DPDNet). Specifically, to improve the performance of detection-based approaches for dense/tiny heads, we propose a density map guided detection module, which leverages density map to improve the head/non-head classification in detection network where the density implies the probability of a pixel being a head, and a depth-adaptive kernel that considers the variances in head sizes is also introduced to generate high-fidelity density map for more robust density map regression. In order to prevent dense heads from being filtered out during post-processing, we utilize such a density map for post-processing of head detection and propose a density map guided NMS strategy. Meanwhile, to improve the ability of detecting small heads, we also propose a depth-guided detection module to generate a dynamic dilated convolution to extract features of heads of different scales, and a depth-aware anchor is further designed for better initialization of anchor sizes in the detection framework. Then we use the bounding boxes whose sizes are generated with depth to train our DPDNet. Considering that existing RGB-D datasets are too small and not suitable for performance evaluation of data-driven based approaches, we collect two large-scale RGB-D crowd counting datasets, which comprise a synthetic dataset and a real-world dataset, respectively. Since the depth value at long-distance positions cannot be obtained in the real-world dataset, we further propose a depth completion method with meta learning, which fully utilizes the synthetic depth data to complete the depth value at long-distance positions. Extensive experiments on our proposed two RGB-D datasets and the MICC RGB-D counting dataset show that our method achieves the best performance for RGB-D crowd counting and localization. Further, our method can be easily extended to RGB image based crowd counting and achieves comparable or even better performance on the RGB datasets for both head counting and localization.

Hidden in plain sight: Integrated population models to resolve partially observable latent population structure

AbstractPopulation models often require detailed information on sex‐, age‐, or size‐specific abundances, but population monitoring programs cannot always acquire data at the desired resolution. Thus, state uncertainty in monitoring data can potentially limit the demographic resolution of management decisions, which may be particularly problematic for stage‐ or size‐structured species subject to consumptive use. American alligators (Alligator mississippiensis; hereafter alligator) have a complex life history characterized by delayed maturity and slow somatic growth, which makes the species particularly sensitive to overharvest. Though alligator populations are subject to recreational harvest throughout their range, the most widely used monitoring method (nightlight surveys) is often unable to obtain size class‐specific counts, which limits the ability of managers to evaluate the effects of harvest policies. We constructed a Bayesian integrated population model (IPM) for alligators in Georgetown County, SC, USA, using records of mark–recapture–recovery, clutch size, harvest, and nightlight survey counts collected locally, and auxiliary information on fecundity, sex ratio, and somatic growth from other studies. We created a multistate mark–recapture–recovery model with six size classes to estimate survival probability, and we linked it to a state‐space count model to derive estimates of size class‐specific detection probability and abundance. Because we worked from a count dataset in which 60% of the original observations were of unknown size, we treated size class as a latent property of detections and developed a novel observation model to make use of information where size could be partly observed. Detection probability was positively associated with alligator size and water temperature, and negatively influenced by water level. Survival probability was lowest in the smallest size class but was relatively similar among the other five size classes (&gt;0.90 for each). While the two nightlight survey count sites exhibited relatively stable population trends, we detected substantially different patterns in size class‐specific abundance and trends between each site, including 30%–50% declines in the largest size classes at the site with greater harvest pressure. Here, we illustrate the use of IPMs to produce high‐resolution output of latent population structure that is partially observed during the monitoring process.

DEFNet: Dual-Branch Enhanced Feature Fusion Network for RGB-T Crowd Counting

Most existing crowd counting approaches use limited information of RGB (red–green–blue) images and fail to suitably extract potential pedestrians in unconstrained scenarios. Moreover, complementary depth maps do not provide information of locations where people are more likely to be present. However, by incorporating optical and thermal information, the recognition of pedestrians may be enhanced considerably. In fact, thermal imaging information is robust to weather and lighting scenarios, and information from targets can be extracted even at nighttime. By combining RGB and thermal imaging information, we propose a dual-branch enhanced feature fusion network (DEFNet) for RGB-T (RGB and thermal) crowd counting. In DEFNet, an intensive data-enhancement module fuses complementary features of the same sizes from the RGB and thermal modalities, thus combining various rich receptive fields and generating powerful fused RGB-T features. These features describe both spatial structures and appearance details, highlighting information of crowd location. Then, an efficient dilation fusion module applies convolutions to the RGB -T features to obtain flexible and specific features, effectively eliminating the influence of background on the crowd information for density map prediction. Finally, high- and low-level features are used to efficiently obtain density maps through a fusion decoding module. Experimental results on an RGB-T crowd counting dataset indicate that the proposed DEFNet outperforms existing approaches. Furthermore, DEFNet can be generalized to handle RGB and depth data.