Multi-class Vehicle Counting System for Multi-view Traffic Videos

We attempt to connect the data from complementary views, i.e., top view from drone-mounted cameras in the air, and side view from wearable cameras on the ground. Collaborative analysis of such complementary-view data can facilitate to build the air-ground cooperative visual system for various kinds of applications. This is a very challenging problem due to the large view difference between top and side views. In this paper, we develop a new approach that can simultaneously handle three tasks: i) localizing the side-view camera in the top view; ii) estimating the view direction of the side-view camera; iii) detecting and associating the same subjects on the ground across the complementary views. Our main idea is to explore the spatial position layout of the subjects in two views. In particular, we propose a spatial-aware position representation method to embed the spatial-position distribution of the subjects in different views. We further design a cross-view video collaboration framework composed of a camera identification module and a subject association module to simultaneously perform the above three tasks. We collect a new synthetic dataset consisting of top-view and side-view video sequence pairs for performance evaluation and the experimental results show the effectiveness of the proposed method.

Pagoda[5]arene (P5), which is composed of five 2,6-dimethoxylanthracene (2,6-DMA) subunits bridged by methylene groups at 1,5 positions, was conveniently synthesized in 43% by trifluoroacetic acid ...

Impaired walking ability and leg health are commonly seen in broilers and can negatively impact their welfare. Commonly, walking ability and leg health are assessed manually, but this is time consuming and can be subjective. Automated approaches for scoring walking ability and leg health at the individual level could therefore have great added value. Here, we studied whether automatically extracted top-view walking features of broilers can be used as a proxy for walking ability and leg health. Top-view videos were collected of unmarked broilers walking through a corridor that was placed inside the home pen. From these videos, four top-view features were derived: 1) lateral body oscillation, calculated as deviations from the movement trajectory of the bird, 2) step count, 3) completion time, and 4) length-width ratio of the virtual bounding box encapsulating the bird while walking as an indicator of wing support. We assessed the relationship of these computer vision-based features with manual gait and leg health scores, including hock burn (HB) and footpad dermatitis (FPD). We observed that birds with worse gait scores (GS) had longer completion times, higher step counts and a trend for higher lateral body oscillation levels in the walkway setup. Unsupervised clustering using the K-means algorithm with these walking features showed potential to distinguish birds with GS3+, although differentiating between GS1 and GS2 proved more challenging. We concluded that the length-width ratio of the bounding box during walking was not a suitable proxy for poor gait. We found no relationship between top-view walking features and mild cases of HB and FPD in broilers. Overall, the results of this study indicate that top-view video recordings can provide insight into birds' walking ability, using features related to movement speed, step count and lateral body oscillation, making automated scoring more feasible on a larger scale in practice. However, these top view features provide little information about mild HB and FPD.

Nowadays, 5G profoundly impacts video surveillance and monitoring services by processing video streams at high-speed with high-reliability, high bandwidth, and secure network connectivity. It also enhances artificial intelligence, machine learning, and deep learning techniques, which require intense processing to deliver near-real-time solutions. In video surveillance, person tracking is a crucial task due to the deformable nature of the human body, various environmental components such as occlusion, illumination, and background conditions, specifically, from a top view perspective where the person’s visual appearance is significantly different from a frontal or side view. In this work, multiple people tracking framework is presented, which uses 5G infrastructure. A top view perspective is used, which offers broad coverage of the scene or field of view. To perform a person tracking deep learning-based tracking by detection framework is proposed, which includes detection by YOLOv3 and tracking by Deep SORT algorithm. Although the model is pre-trained using the frontal view images, even then, it gives good detection results. In order to further enhance the accuracy of the detection model, the transfer learning approach is adopted. In this way, a detection model takes advantage of a pre-trained model appended with an additional trained layer using top view data set. To evaluate the performance, experiments are carried out on different top view video sequences. Experimental results reveal that transfer learning improves the overall performance, detection accuracy, and reduces false positives. The deep learning detection model YOLOv3 achieves detection accuracy of 92% with a pre-trained model without transfer learning and 95% with transfer learning. The tracking algorithm Deep SORT also achieves excellent results with a tracking accuracy of 96%.

π-Pimer, π-Dimer, π-Trimer, and 1D π-Stacks in a Series of Benzene Triimide Radical Anions: Substituent-Modulated π Interactions and Physical Properties in Crystalline State

<div class="htmlview paragraph">A pictorial reference for the identification and location of the filaments and electrical contacts for common motor vehicle dual filament sealed beam headlamps and dual filament replaceable headlamp bulbs is presented. The filaments and contacts shown in the reference photographs are clearly identified to facilitate comparison to headlamps or bulbs under investigation.</div>

Nighttime vehicle detection is part of an intelligent transportation system for road safety, driving navigation, and surveillance at night. However, previous nighttime vehicle detection methods only deal with a single class or two classes of vehicles. This paper presents an effective system based on cascade feature selection and a coarse-to-fine mechanism for detecting preceding multiclass vehicles at night. First, we train a coarse-level classifier using contrast features. Second, we combine a cascade selection framework with feature mutual correlation, similarity and regions’ overlap to select a set of image regions, where we can extract effective features. Three features, including local binary pattern, histogram of oriented gradients, and four direction features, are extracted from the selected regions for training a fine-level multiclass vehicle classifier. During the detection stage, we utilize a coarse-to-fine mechanism. In the coarse level, a multiscale sliding window is classified by the coarse-level classifier to find regions of interest (ROIs) that are likely to be vehicles. In the fine level, these ROIs are identified by the trained multiclass vehicle classifier. Evaluations on a Hong Kong nighttime multiclass vehicle dataset show that our proposed system successfully detects a car, taxi, bus, and minibus within nighttime images under different scenes. Quantitatively, our proposed system obtains 95.48% detection rate at 0.055 false positives per image, outperforming some state-of-the-art detection approaches including two current nighttime vehicle detection methods, and two deep learning-based methods.

It is important to disclose the wake flow around cylinders for understanding the mechanisms of heat transfer enhancement in flow networks. An experimental study was performed to visualize the wake flow in the gap between two rectangular cylinders. Square and rectangular cylinders are selected at Reynolds number range of less than 600 in the experiment. The wake flows are visualized at a top view, side views and front views using the fluorescein dye illuminated by a laser light sheet. Different shapes of vortices are shown in the wake flows visualized at three views. Flow visualization allows a better understanding of vortex structures of flow in the gap between two rectangular cylinders.

BackgroundIn the last few years high-throughput analysis methods have become state-of-the-art in the life sciences. One of the latest developments is automated greenhouse systems for high-throughput plant phenotyping. Such systems allow the non-destructive screening of plants over a period of time by means of image acquisition techniques. During such screening different images of each plant are recorded and must be analysed by applying sophisticated image analysis algorithms.ResultsThis paper presents an image analysis pipeline (HTPheno) for high-throughput plant phenotyping. HTPheno is implemented as a plugin for ImageJ, an open source image processing software. It provides the possibility to analyse colour images of plants which are taken in two different views (top view and side view) during a screening. Within the analysis different phenotypical parameters for each plant such as height, width and projected shoot area of the plants are calculated for the duration of the screening. HTPheno is applied to analyse two barley cultivars.ConclusionsHTPheno, an open source image analysis pipeline, supplies a flexible and adaptable ImageJ plugin which can be used for automated image analysis in high-throughput plant phenotyping and therefore to derive new biological insights, such as determination of fitness.

<strong class="journal-contentHeaderColor">Abstract.</strong> With the advancement of computer vision, artificial intelligence, and remote sensing technologies, deep learning algorithms are increasingly used in terrestrial, airborne, and spaceborne-based fire detection systems. The performance and generalization of these data-driven fire detection algorithms, however, are restricted by the limited number and source of fire detection datasets. A large-scale fire detection benchmark dataset covering complex and varied fire scenarios is urgently needed. This work constructs a 100,000-level Flame and Smoke Detection Dataset (FASDD) based on multi-source heterogeneous flame and smoke images. It holds rich variations in image size, resolution, illumination (day and night), scenario (indoor and outdoor), image range (far and near), viewing angle (top view and side view), platform (surveillance cameras, drones, and satellites), and data source (Internet, social media, and open-access fire datasets). To the best of our knowledge, FASDD is currently the most versatile and comprehensive dataset for fire detection. It provides a challenging benchmark to drive the continuous evolution of fire detection models. Additionally, we formulate a unified workflow for preprocessing, annotation, and quality control of fire samples. Out-of-the-box annotations are published in four different formats for training deep learning models. Extensive performance evaluations based on classical methods show that most of the object detection models trained on FASDD can achieve satisfactory fire detection results, and especially YOLOv5x achieves nearly 80 % mAP@0.5 accuracy on heterogeneous images spanning two domains of computer vision and remote sensing. And the application in wildfire location demonstrates that deep learning models trained on our dataset can be used in recognizing and monitoring forest fires. Deep learning models trained with FASDD can be simultaneously deployed on satellites, drones, and ground sensors, thus realizing collaborative fire observation and detection in a space-air-ground integrated network environment. The dataset is available from the Science Data Bank website at <a href="https://doi.org/10.57760/sciencedb.j00104.00103" target="_blank" rel="noopener">https://doi.org/10.57760/sciencedb.j00104.00103</a> (Wang et al., 2022a).

In the human diet, particularly for most of the vegetarian population, mungbean (Vigna radiata L. Wilczek) is an inexpensive and environmentally friendly source of protein. Being a short-duration crop, mungbean fits well into different cropping systems dominated by staple food crops such as rice and wheat. Hence, knowing the growth and production pattern of this important legume under various soil moisture conditions gains paramount significance. Toward that end, 24 elite mungbean genotypes were grown with and without water stress for 25 days in a controlled environment. Top view and side view (two) images of all genotypes captured by a high-resolution camera installed in the high-throughput phenomics were analyzed to extract the pertinent parameters associated with plant features. We tested eight different multivariate models employing machine learning algorithms to predict fresh biomass from different features extracted from the images of diverse genotypes in the presence and absence of soil moisture stress. Based on the mean absolute error (MAE), root mean square error (RMSE), and R squared (R2) values, which are used to assess the precision of a model, the partial least square (PLS) method among the eight models was selected for the prediction of biomass. The predicted biomass was used to compute the plant growth rates and water-use indices, which were found to be highly promising surrogate traits as they could differentiate the response of genotypes to soil moisture stress more effectively. To the best of our knowledge, this is perhaps the first report stating the use of a phenomics method as a promising tool for assessing growth rates and also the productive use of water in mungbean crop.

Although evolution dynamic investigation onto electrochemically generated bubbles on electrode surfaces is crucial for releasing “bubble blocking” effects, traditional “top view” or “side view” methods make spatiotemporal interfacial revolutions difficult to achieve, which are critical for a deeper understanding and better electrode design. In this study, a novel method utilizing surface plasmon resonance (SPR) is proposed to probe bubble interfacial dynamics. By monitoring reflectivity changes and utilizing differential imaging, real‐time fluctuations in the contact area can be accurately visualized. The investigation onto bubble growth events on both aerophilic and aerophobic surfaces reveals that they occur in very distinct patterns. Due to the “necking effect”, the reflectivity of aerophilic electrodes rushed to its highest value and remained or gradually decreased, followed by a sudden drop due to bubble disengagement. While the aerophobic ones increased to the bubble release threshold after a downward parabola function, then dropped abruptly without plateau. This in situ imaging technique provides a feasible tool for bubble disengagement dynamic investigations on electrode surface. Based on current observations, it is revealed that changing the wettability of electrodes causes substantial changes in bubble evolution dynamics.

We present a fully-automatic image processing pipeline to build a set of 2D morphable models of three craniofacial profiles from orthogonal viewpoints, side view, front view and top view, using a set of 3D head surface images. Subjects in this dataset wear a close-fitting latex cap to reveal the overall skull shape. Texture-based 3D pose normalization and facial landmarking are applied to extract the profiles from 3D raw scans. Fully-automatic profile annotation, subdivision and registration methods are used to establish dense correspondence among sagittal profiles. The collection of sagittal profiles in dense correspondence are scaled and aligned using Generalised Procrustes Analysis (GPA), before applying principal component analysis to generate a morphable model. Additionally, we propose a new alternative alignment called the Ellipse Centre Nasion (ECN) method. Our model is used in a case study of craniosynostosis intervention outcome evaluation, and the evaluation reveals that the proposed model achieves state-of-the-art results. We make publicly available both the morphable models and the profile dataset used to construct it.

A preliminary investigation of the post heat treatment for Ti6Al4V tensile specimens fabricated via selective laser melting (SLM) are reported. The resultant microstructure and mechanical behavior at room temperature, normal and parallel to the building direction, was investigated in detail. Moreover, the as-prepared samples were compared with the samples post heat treated at 600°C~800°C for 2 h. It was found that the SLM fabricated samples consisted of long columnar original β grains together with parallel acicular α' martensitic structure dominated in the β matrix, which was observed in the side view, and fully equaixed β grains pattern in the top view. The post heat treatment did not completely disrupt the layered structure, but leading to the acicular α' martensites decomposing into α platelets distributed in a metastable β matrix. Ultimate tensile strength and elongation were analyzed and explained based on the microstructure evolution. After post heat treatment, an increase in elongation was observed for the tensile samples, inducing more ductile like morphology after fracture.

We present photoluminescence and resonant Raman measurements of atomically thin layers of molybdenum disulfide (MoS2) in the energy range of the A and B optical transitions. We found a Stokes shift of the transition energies in the resonant Raman measurements compared to the photoluminescence. The Stokes shift increases for the bilayer compared to the single layer. Furthermore, in resonance with the A transition, the Raman intensity of the A1g mode of the bilayer is six times larger than that of the single layer. magnified imageSchematic crystal structure of molybdenum disulfide. (a) Top view (slightly tilted). (b) Side view.