Putative volatile biomarkers of bovine tuberculosis infection in breath, skin and feces of cattle.

Bovine tuberculosis (bTB) is an infectious disease with significant impact on animal health, public health and international trade. Standard bTB screening in live cattle consists in injecting tuberculin and measuring the swelling at the place of injection few days later. This procedure is expensive, time-consuming, logistically challenging, and is not conclusive before performing confirmatory tests and additional analysis. The analysis of the volatile organic compounds (VOCs) emitted by non-invasive biological samples can provide an alternative diagnostic approach suitable for bTB screening. In the present study, we analyzed VOC samples emitted through the breath, feces and skin of 18 cows diagnosed with bTB from three farms from Romania, as well as of 27 negative cows for bTB from the same farms. Analytical studies employing gas chromatography coupled to mass spectrometry revealed 80 VOCs emitted through the breath, 200 VOCs released by feces, and 80 VOCs emitted through the skin. Statistical analysis of these compounds allowed the identification of 3 tentative breath VOC biomarkers (acetone; 4-methyldecane; D-limonene), 9 tentative feces VOC biomarkers (toluene; [(1,1-dimethylethyl)thio]acetic acid; alpha-thujene; camphene; phenol; o-cymene; 3-(1,1-dimethylethyl)-2,2,4,4-tetramethyl-3-pentanol; 2,5-dimethylhexane-2,5-dihydroperoxide; 2,4-di-tert-butylphenol), and 3 tentative skin VOC biomarkers (ammonia; 1-methoxy-2-propanol; toluene). The possible pathway of these volatile biomarkers is discussed.

Finding a Suitable Class Distribution for Building Histological Images Datasets Used in Deep Model Training-The Case of Cancer Detection.

The class distribution of a training dataset is an important factor which influences the performance of a deep learning-based system. Understanding the optimal class distribution is therefore crucial when building a new training set which may be costly to annotate. This is the case for histological images used in cancer diagnosis where image annotation requires domain experts. In this paper, we tackle the problem of finding the optimal class distribution of a training set to be able to train an optimal model that detects cancer in histological images. We formulate several hypotheses which are then tested in scores of experiments with hundreds of trials. The experiments have been designed to account for both segmentation and classification frameworks with various class distributions in the training set, such as natural, balanced, over-represented cancer, and over-represented non-cancer. In the case of cancer detection, the experiments show several important results: (a) the natural class distribution produces more accurate results than the artificially generated balanced distribution; (b) the over-representation of non-cancer/negative classes (healthy tissue and/or background classes) compared to cancer/positive classes reduces the number of samples which are falsely predicted as cancer (false positive); (c) the least expensive to annotate non-ROI (non-region-of-interest) data can be useful in compensating for the performance loss in the system due to a shortage of expensive to annotate ROI data; (d) the multi-label examples are more useful than the single-label ones to train a segmentation model; and (e) when the classification model is tuned with a balanced validation set, it is less affected than the segmentation model by the class distribution of the training set.

Non-Invasive Method to Detect Infection with Mycobacterium tuberculosis Complex in Wild Boar by Measurement of Volatile Organic Compounds Obtained from Feces with an Electronic Nose System.

More effective methods to detect bovine tuberculosis, caused by Mycobacterium bovis, in wildlife, is of paramount importance for preventing disease spread to other wild animals, livestock, and human beings. In this study, we analyzed the volatile organic compounds emitted by fecal samples collected from free-ranging wild boar captured in Doñana National Park, Spain, with an electronic nose system based on organically-functionalized gold nanoparticles. The animals were separated by the age group for performing the analysis. Adult (>24 months) and sub-adult (12-24 months) animals were anesthetized before sample collection, whereas the juvenile (<12 months) animals were manually restrained while collecting the sample. Good accuracy was obtained for the adult and sub-adult classification models: 100% during the training phase and 88.9% during the testing phase for the adult animals, and 100% during both the training and testing phase for the sub-adult animals, respectively. The results obtained could be important for the further development of a non-invasive and less expensive detection method of bovine tuberculosis in wildlife populations.

Accurate and Efficient Intracranial Hemorrhage Detection and Subtype Classification in 3D CT Scans with Convolutional and Long Short-Term Memory Neural Networks.

In this paper, we present our system for the RSNA Intracranial Hemorrhage Detection challenge, which is based on the RSNA 2019 Brain CT Hemorrhage dataset. The proposed system is based on a lightweight deep neural network architecture composed of a convolutional neural network (CNN) that takes as input individual CT slices, and a Long Short-Term Memory (LSTM) network that takes as input multiple feature embeddings provided by the CNN. For efficient processing, we consider various feature selection methods to produce a subset of useful CNN features for the LSTM. Furthermore, we reduce the CT slices by a factor of 2×, which enables us to train the model faster. Even if our model is designed to balance speed and accuracy, we report a weighted mean log loss of 0.04989 on the final test set, which places us in the top 30 ranking (2%) from a total of 1345 participants. While our computing infrastructure does not allow it, processing CT slices at their original scale is likely to improve performance. In order to enable others to reproduce our results, we provide our code as open source. After the challenge, we conducted a subjective intracranial hemorrhage detection assessment by radiologists, indicating that the performance of our deep model is on par with that of doctors specialized in reading CT scans. Another contribution of our work is to integrate Grad-CAM visualizations in our system, providing useful explanations for its predictions. We therefore consider our system as a viable option when a fast diagnosis or a second opinion on intracranial hemorrhage detection are needed.

Assessment of Electronic Sensing Techniques for the Rapid Identification of Alveolar Echinococcosis through Exhaled Breath Analysis.

Here we present a proof-of-concept study showing the potential of a chemical gas sensors system to identify the patients with alveolar echinococcosis disease through exhaled breath analysis. The sensors system employed comprised an array of three commercial gas sensors and a custom gas sensor based on WO3 nanowires doped with gold nanoparticles, optimized for the measurement of common breath volatile organic compounds. The measurement setup was designed for the concomitant measurement of both sensors DC resistance and AC fluctuations during breath samples exposure. Discriminant Function Analysis classification models were built with features extracted from sensors responses, and the discrimination of alveolar echinococcosis was estimated through bootstrap validation. The commercial sensor that detects gases such as alkane derivatives and ethanol, associated with lipid peroxidation and intestinal gut flora, provided the best classification (63.4% success rate, 66.3% sensitivity and 54.6% specificity) when sensors' responses were individually analyzed, while the model built with the AC features extracted from the responses of the cross-reactive sensors array yielded 90.2% classification success rate, 93.6% sensitivity and 79.4% specificity. This result paves the way for the development of a noninvasive, easy to use, fast and inexpensive diagnostic test for alveolar echinococcosis diagnosis at an early stage, when curative treatment can be applied to the patients.

Diagnosis of Human Echinococcosis via Exhaled Breath Analysis: A Promise for Rapid Diagnosis of Infectious Diseases Caused by Helminths.

Background: Human echinococcosis is a neglected infectious disease affecting more than 1 million people globally. Its diagnosis is expensive and difficult because of lack of adequate resources in low-resource locations, where most cases occur. Methods: A group of volunteers diagnosed with the 2 main types of echinococcosis and corresponding control groups were recruited from hospitals in Tunisia (32 patients with cystic echinococcosis and 43 controls) and Poland (16 patients with alveolar echinococcosis and 8 controls). Breath samples were collected from all patients and analyzed by gas chromatography coupled to mass spectrometry, and a specifically developed electronic nose system. Results: The chemical analysis revealed statistically different concentrations of 2 compounds in the breath of patients with cystic echinococcosis compared to controls, and statistically different concentrations of 7 compounds in the breath of patients with alveolar echinococcosis compared to controls. The discrimination accuracy achieved by the electronic nose system was 100% for cystic echinococcosis and 92.9% for alveolar echinococcosis, while the discrimination accuracy between these 2 patient groups was 92.1%. Conclusion: Here we advocate a noninvasive, fast, easy-to-operate and nonexpensive diagnostic tool for the diagnosis of human echinococcosis disease through exhaled breath analysis, suitable for early diagnosis and population screening.

Intra-Areolar Pexy: The "Compass Rose" Suture Technique for Small and Moderate Areola Herniation.

BACKGROUND: Common methods for addressing minor tuberous breast (type 0; normal mammary base breast with isolated areolar herniation) do not solve the underlying pathology and may leave an unsightly scar. OBJECTIVES: To assess the efficacy and safety of a simple, scarless method based on a novel "compass rose" technique, which addresses the etiology of minor tuberous breast and other cases desiring small areolar diameter reductions. METHODS: The technique uses 3 layers of suturing, 1 in a compass pattern and 2 in round block. Retrospective data are provided for 77 consecutive women (141 breasts) undergoing cosmetic breast augmentation surgery and requiring unilateral or bilateral type 0 tuberous breast correction (n = 22) or areolar width reduction ≤ 15 mm (n = 55). RESULTS: Baseline mean age and body mass index were 29.7 ± 6.0 years and 19.4 ± 1.5 kg/m2, respectively. Patients were followed up for a mean of 27.2 ± 19.5 months. In tuberous breast, mean Northwood Index decreased from 0.55 ± 0.06 at baseline to 0.36 ± 0.02 at ≥ 6 months post surgery, indicating minimal residual deformity. In areolar width reductions without tuberosity, mean width decreases of 11.3 ± 1.8 mm (20.0%) were achieved. Patient satisfaction was high with regard to lack of scarring, stability of the result, and overall breast attractiveness. Eight complications were recorded (n = 2 superficial hematoma; n = 3 hypersensibility; n = 2 knot palpability; n = 1 recurrence); all were resolved. CONCLUSIONS: The method is effective and safe, and may be applicable across patients with type 0 tuberous breast or desiring small areolar diameter reductions.

Fabrication and characterisation of ligand-functionalised ultrapure monodispersed metal nanoparticle nanoassemblies employing advanced gas deposition technique.

Here, we report for the first time the fabrication of ligand-functionalised ultrapure monodispersed metal nanoparticles (Au, Cu, and Pt) from their pure metal precursors using the advanced gas deposition technique. The experimental conditions during nanoparticle formation were adjusted in order to obtain ultrafine isolated nanoparticles on different substrates. The morphology and surface analysis of the as-deposited metal nanoparticles were investigated using scanning electron microscopy, x-ray diffraction and Fourier transform infra-red spectroscopy, which demonstrated the formation of highly ordered pure crystalline nanoparticles with a relatively uniform size distribution of ∼10 nm (Au), ∼4 nm (Cu) and ∼3 nm (Pt), respectively. A broad range of organic ligands containing thiol or amine functional groups were attached to the nanoparticles to form continuous networks of nanoparticle-ligand nanoassemblies, which were characterised by scanning electron microscopy and x-ray photoelectron spectroscopy. The electrical resistance of the functional nanoassemblies deposited in the gap spacing of two microfabricated parallel Au electrodes patterned on silicon substrates ranged between tens of kΩ and tens of MΩ, which is suitable for use in many applications including (bio)chemical sensors, surface-enhanced Raman spectroscopy and molecular electronic rectifiers.

Insurance, Public Assistance, and Household Flood Risk Reduction: A Comparative Study of Austria, England, and Romania.

In light of increasing losses from floods, many researchers and policymakers are looking for ways to encourage flood risk reduction among communities, business, and households. In this study, we investigate risk-reduction behavior at the household level in three European Union Member States with fundamentally different insurance and compensation schemes. We try to understand if and how insurance and public assistance influence private risk-reduction behavior. Data were collected using a telephone survey (n = 1,849) of household decisionmakers in flood-prone areas. We show that insurance overall is positively associated with private risk-reduction behavior. Warranties, premium discounts, and information provision with respect to risk reduction may be an explanation for this positive relationship in the case of structural measures. Public incentives for risk-reduction measures by means of financial and in-kind support, and particularly through the provision of information, are also associated with enhancing risk reduction. In this study, public compensation is not negatively associated with private risk-reduction behavior. This does not disprove such a relationship, but the negative effect may be mitigated by factors related to respondents' capacity to implement measures or social norms that were not included in the analysis. The data suggest that large-scale flood protection infrastructure creates a sense of security that is associated with a lower level of preparedness. Across the board there is ample room to improve both public and private policies to provide effective incentives for household-level risk reduction.

Non-Gaussian Resistance Fluctuations in Gold-Nanoparticle-Based Gas Sensors: An Appraisal of Different Evaluation Techniques.

Volatile organic compounds, such as formaldehyde, can be used as biomarkers in human exhaled breath in order to non-invasively detect various diseases, and the same compounds are of much interest also in the context of environmental monitoring and protection. Here, we report on a recently-developed gas sensor, based on surface-functionalized gold nanoparticles, which is able to generate voltage noise with a distinctly non-Gaussian component upon exposure to formaldehyde with concentrations on the ppm level, whereas this component is absent, or at least much weaker, when the sensor is exposed to ethanol or to pure air. We survey four different statistical methods to elucidate a non-Gaussian component and assess their pros and cons with regard to efficient gas detection. Specifically, the non-Gaussian component was clearly exposed in analysis using level-crossing parameters, which require nothing but a modest computational effort and simple electronic circuitry, and analogous results could be reached through the bispectrum function, albeit with more intense computation. Useful information could be obtained also via the Lévy-stable distribution and, possibly, the second spectrum.

Assessment of ovarian cancer conditions from exhaled breath.

We present a pilot study that aims to examine the possibility to easily and noninvasively detect and discriminate females with ovarian cancer (OC) from females that have no tumor(s) and from females that have benign genital tract neoplasia, using exhaled breath samples. The study is based on clinical samples and data from 182 females, as follows: 48 females with OC, 48 tumor-free controls and 86 females with benign gynecological neoplasia. Analysis of the breath samples with gas chromatography linked with mass spectrometry shows that decanal, nonanal, styrene, 2-butanone and hexadecane could serve as potential volatile markers for OC. Analysis of the same samples with tailor-made nanoarrays shows good discrimination between females with OC and females that have either no tumor or benign genital tract neoplasia (71% for accuracy, sensitivity and specificity). Conversely, the nanoarray output shows excellent discrimination between the OC patients and the tumor-free controls (79% sensitivity, 100% specificity and 89% accuracy). These results suggest that the nanoarray approach might be useful to avoid unnecessary complicated or expensive tests for tumor-free females in case of a negative result. In the case of positive result, the test will indicate with high probability the presence of OC.

Prediction of Subclinical and Clinical Multiple Organ Failure Dysfunction in Breast Cancer Patients-A Review Using AI Tools.

This review explores the interconnection between precursor lesions of breast cancer (typical ductal hyperplasia, atypical ductal/lobular hyperplasia) and the subclinical of multiple organ failure syndrome, both representing early stages marked by alterations preceding clinical symptoms, undetectable through conventional diagnostic methods. Addressing the question "Why patients with breast cancer exhibit a tendency to deteriorate", this study investigates the biological progression from a subclinical multiple organ failure syndrome, characterized by insidious but indisputable lesions, to an acute (clinical) state resembling a cascade akin to a waterfall or domino effect, often culminating in the patient's demise. A comprehensive literature search was conducted using PubMed, Google Scholar, and Scopus databases in October 2023, employing keywords such as "MODS", "SIRS", "sepsis", "pathophysiology of MODS", "MODS in cancer patients", "multiple organ failure", "risk factors", "cancer", "ICU", "quality of life", and "breast cancer". Supplementary references were extracted from the retrieved articles. This study emphasizes the importance of early identification and prevention of the multiple organ failure cascade at the inception of the malignant state, aiming to enhance the quality of life and extend survival. This pursuit contributes to a deeper understanding of risk factors and viable therapeutic options. Despite the existence of the subclinical multiple organ failure syndrome, current diagnostic methodologies remain inadequate, prompting consideration of AI as an increasingly crucial tool for early identification in the diagnostic process.

Volatile fingerprints of cancer specific genetic mutations.

We report on a new concept for profiling genetic mutations of (lung) cancer cells, based on the detection of patterns of volatile organic compounds (VOCs) emitted from cell membranes, using an array of nanomaterial-based sensors. In this in-vitro pilot study we have derived a volatile fingerprint assay for representative genetic mutations in cancer cells that are known to be associated with targeted cancer therapy. Five VOCs were associated with the studied oncogenes, using complementary chemical analysis, and were discussed in terms of possible metabolic pathways. The reported approach could lead to the development of novel methods for guiding treatments, so that patients could benefit from safer, more timely and effective interventions that improve survival and quality of life while avoiding unnecessary invasive procedures. Studying clinical samples (tissue/blood/breath) will be required as next step in order to determine whether this cell-line study can be translated into a clinically useful tool. FROM THE CLINICAL EDITOR: In this novel study, a new concept for profiling genetic mutations of (lung) cancer cells is described, based on the detection of patterns of volatile organic compounds emitted from cell membranes, using an array of nano-gold based sensors.

Diffusion Models in Vision: A Survey.

Denoising diffusion models represent a recent emerging topic in computer vision, demonstrating remarkable results in the area of generative modeling. A diffusion model is a deep generative model that is based on two stages, a forward diffusion stage and a reverse diffusion stage. In the forward diffusion stage, the input data is gradually perturbed over several steps by adding Gaussian noise. In the reverse stage, a model is tasked at recovering the original input data by learning to gradually reverse the diffusion process, step by step. Diffusion models are widely appreciated for the quality and diversity of the generated samples, despite their known computational burdens, i.e., low speeds due to the high number of steps involved during sampling. In this survey, we provide a comprehensive review of articles on denoising diffusion models applied in vision, comprising both theoretical and practical contributions in the field. First, we identify and present three generic diffusion modeling frameworks, which are based on denoising diffusion probabilistic models, noise conditioned score networks, and stochastic differential equations. We further discuss the relations between diffusion models and other deep generative models, including variational auto-encoders, generative adversarial networks, energy-based models, autoregressive models and normalizing flows. Then, we introduce a multi-perspective categorization of diffusion models applied in computer vision. Finally, we illustrate the current limitations of diffusion models and envision some interesting directions for future research.

Detection of Counterfeit Perfumes by Using GC-MS Technique and Electronic Nose System Combined with Chemometric Tools.

The Scientific Committee on Cosmetic and Non-Food Products has identified 26 compounds that may cause contact allergy in consumers when present in concentrations above certain legal thresholds in a product. Twenty-four of these compounds are volatiles and can be analyzed by gas chromatography-mass spectrometry (GC-MS) or electronic nose (e-nose) technologies. This manuscript first describes the use of the GC-MS approach to identify the main volatile compounds present in the original perfumes and their counterfeit samples. The second part of this work focusses on the ability of an e-nose system to discriminate between the original fragrances and their counterfeits. The analyses were carried out using the headspace of the aqueous solutions. GC-MS analysis revealed the identification of 10 allergens in the perfume samples, some of which were only found in the imitated fragrances. The e-nose system achieved a fair discrimination between most of the fragrances analyzed, with the counterfeit fragrances being clearly separated from the original perfumes. It is shown that associating the e-nose system to the appropriate classifier successfully solved the classification task. With Principal Component Analysis (PCA), the three first principal components represented 98.09% of the information in the database.

Challenges during the realization of an international research project on leishmaniasis in Colombia.

Leishmaniasis is an infectious disease that belongs to the top 10 neglected tropical diseases. It mainly affects the poor population from tropical and subtropical areas of the World, which lacks sufficient resources and means to fight against this disease. With this in mind, the European Commission has funded an international collaborative research project in which are participating various institutions from South America, North Africa and Europe. The main objective of this project is the development of a fast, less expensive, non-invasive and easy to use alternative method for leishmaniasis diagnosis in dogs, one of the main reservoirs of leishmaniasis spread to humans. In this perspective article, we present our personal insight and opinion regarding the challenges of realizing a joint international research project on leishmaniasis in Colombia, a country where leishmaniasis is endemic, as well as regarding the involvement of the Public Health institutions and the local population from this country.

Dorsal nasal augmentation with "open sandwich" graft consisting of conchal cartilage and retroauricular fascia.

BACKGROUND: Augmenting the nasal dorsum with conchal grafts can cause visible irregularities over time due to the morphological qualities of that material. OBJECTIVES: This study describes the senior authors' technique of dorsal nasal augmentation with chondrofascial "open sandwich" grafts consisting of pieces of conchal cartilage and retroauricular fascia. The authors assess the efficiency and reliability of this graft in nasal dorsal augmentation. METHODS: The authors retrospectively reviewed the cases of 19 patients who underwent dorsal augmentation rhinoplasty with chondrofascial grafts. Both cartilage and fascia were harvested through the same incision in the retroauricular sulcus. The conchal graft was cut in 2 to 4 pieces and slightly crushed with tissue forceps. The pieces of cartilage were arranged and fixed to the fascia in different patterns according to the nasal dorsum contour. RESULTS: The follow-up period ranged from 12 to 35 months in 16 patients who qualified for inclusion in the final data. The maximal thickness of the chondrofascial graft was 4.5 mm. There were no major complications in the recipient area, except 1 case of undercorrection. There were 2 complications in the donor area. In 1 case, a hematoma was treated conservatively. In another case, a strip of skin necrosis in the conchal area occurred and was treated by excision and direct suture with satisfactory resolution. CONCLUSIONS: The autologous chondrofascial graft is appropriate for slight to mild dorsal nasal augmentation. The method, as with most rhinoplasty techniques, requires careful and judicious preoperative examination, planning, and execution. The postoperative scar is inconspicuous and the donor site morbidity is minimal.

Teacher-student training and triplet loss to reduce the effect of drastic face occlusion: Application to emotion recognition, gender identification and age estimation.

We study a series of recognition tasks in two realistic scenarios requiring the analysis of faces under strong occlusion. On the one hand, we aim to recognize facial expressions of people wearing virtual reality headsets. On the other hand, we aim to estimate the age and identify the gender of people wearing surgical masks. For all these tasks, the common ground is that half of the face is occluded. In this challenging setting, we show that convolutional neural networks trained on fully visible faces exhibit very low performance levels. While fine-tuning the deep learning models on occluded faces is extremely useful, we show that additional performance gains can be obtained by distilling knowledge from models trained on fully visible faces. To this end, we study two knowledge distillation methods, one based on teacher-student training and one based on triplet loss. Our main contribution consists in a novel approach for knowledge distillation based on triplet loss, which generalizes across models and tasks. Furthermore, we consider combining distilled models learned through conventional teacher-student training or through our novel teacher-student training based on triplet loss. We provide empirical evidence showing that, in most cases, both individual and combined knowledge distillation methods bring statistically significant performance improvements. We conduct experiments with three different neural models (VGG-f, VGG-face and ResNet-50) on various tasks (facial expression recognition, gender recognition, age estimation), showing consistent improvements regardless of the model or task.

Non-Invasive Staging of In Vitro Mice Embryos by Means of Volatolomics.

Current methods for embryo selection are limited. This study assessed a novel method for the prediction of embryo developmental potential based on the analysis of volatile organic compounds (VOCs) emitted by embryo samples. The study included mice embryos monitored during the pre-implantation period. Four developmental stages of the embryos were tested, covering the period from 1 to 4 days after fecundation. In each stage, the VOCs released by the embryos were collected and examined by employing two different volatolomic techniques, gas-chromatography coupled to mass-spectrometry (GC-MS) and a nanoarray of chemical gas sensors. The GC-MS study revealed that the VOC patterns emanating from embryo samples had statistically different values at different stages of embryo development. The sensor nanoarray was capable of classifying the developmental stages of the embryos. The proposed volatolomics analysis approach for embryos presents a promising potential for predicting their developmental stage. In combination with conventional morphokinetic parameters, it could be effective as a predictive model for detecting metabolic shifts that affect embryo quality before preimplantation.

A Background-Agnostic Framework With Adversarial Training for Abnormal Event Detection in Video.

Abnormal event detection in video is a complex computer vision problem that has attracted significant attention in recent years. The complexity of the task arises from the commonly-adopted definition of an abnormal event, that is, a rarely occurring event that typically depends on the surrounding context. Following the standard formulation of abnormal event detection as outlier detection, we propose a background-agnostic framework that learns from training videos containing only normal events. Our framework is composed of an object detector, a set of appearance and motion auto-encoders, and a set of classifiers. Since our framework only looks at object detections, it can be applied to different scenes, provided that normal events are defined identically across scenes and that the single main factor of variation is the background. This makes our method background agnostic, as we rely strictly on objects that can cause anomalies, and not on the background. To overcome the lack of abnormal data during training, we propose an adversarial learning strategy for the auto-encoders. We create a scene-agnostic set of out-of-domain pseudo-abnormal examples, which are correctly reconstructed by the auto-encoders before applying gradient ascent on the pseudo-abnormal examples. We further utilize the pseudo-abnormal examples to serve as abnormal examples when training appearance-based and motion-based binary classifiers to discriminate between normal and abnormal latent features and reconstructions. Furthermore, to ensure that the auto-encoders focus only on the main object inside each bounding box image, we introduce a branch that learns to segment the main object. We compare our framework with the state-of-the-art methods on four benchmark data sets, using various evaluation metrics. Compared to existing methods, the empirical results indicate that our approach achieves favorable performance on all data sets. In addition, we provide region-based and track-based annotations for two large-scale abnormal event detection data sets from the literature, namely ShanghaiTech and Subway.