Neural Networks For Object Detection Research Articles

Visual Relationship Detection for Workplace Safety Applications

Applications of object and visual relationship detection for safety and security applications are in its infancy. State-of-the-art computer vision research is largely focused on improving mean Average Precision and mean Average Recall performance on standard, general datasets such as the Verbs in COmmon objects in COntext and the Visual Genome dataset and rarely mention the potential of such models in safety and security scenarios. We propose to train and develop an object and visual relationship detection neural network to be used as part of the back-end model for a decision support system. We use a naive Bayesian Network to determine scenarios where our proposed object and visual relationship detection network is error-prone. We also release a graphical user interface which demonstrates how our back-end neural network and naive Bayesian Network can be used for hazardous workplace safety and security applications. All models and source code can be found at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/th-truong/rd-gui</uri> for future research and applications of object and visual relationship detection models in the safety and security sector.

MCCANet: A multispectral class-constraint attentional neural network for object detection in mining scenes

Ion adsorption rare earth mining activities have caused serious ecological problems. It is critical to monitor mining activities through precise object detection. However, in the special scenes of mining sites captured by high-resolution remote sensing images, the large amount of complex background information intensifies the semantic confusion among geographic entities. Therefore, the accurate detection of ground objects in mining areas remains a challenging task. This paper proposes an innovative multispectral class-constraint attentional neural network, referred to as MCCANet, which can use multispectral channel information to derive intra-class and inter-class implicit semantic relations. It forms attention with class information constraints for imposing stronger class supervision in order to suppress the background information and highlight the feature differences of the foreground object in complex scenes. Furthermore, a dataset dedicated to rare earth mining areas is constructed to validate the efficiency of the proposed method. The obtained results show that MCCANet has high accuracy in targeting potential objects within mining scenes while significantly mitigating the high false alarm rates resulting from complex backgrounds. Notably, when compared to current popular deep learning models, MCCANet outperforms others by achieving the highest level of detection accuracy.

Research on Convolution Neural Network for Object Detection and Recognition

The state-of-the-art object detection networks for natural images have recently demonstrated impressive performances. However, the complexity of object’s shape and orientation exposes the limited capacity of these networks for strip-like rotated assembled object detection which are common in any dataset as well as real im- ages. In this project, I embrace this observation and introduce the Faster Rotated Region-based Convolutional Neural Network (Faster RR-CNN), which can learn and accurately extract features of rotated regions and arbitrary-oriented objects precisely. In comparison with the classic Faster RCNN, Faster RR-CNN has three important new components including a skew non-maximum suppression, a rotated bounding box regression model and a rotated region of interest (RRoI) pooling layer. I conduct experiments using the PASCAL VOC 2012 dataset, demonstrat-ing the potential ability of this novel network in detecting oriented objects.

Underwater object detection and temporal signal detection in turbid water using 3D-integral imaging and deep learning.

Underwater scattering caused by suspended particles in the water severely degrades signal detection performance and poses significant challenges to the problem of object detection. This paper introduces an integrated dual-function deep learning-based underwater object detection and classification and temporal signal detection algorithm using three-dimensional (3D) integral imaging (InIm) under degraded conditions. The proposed system is an efficient object classification and temporal signal detection system for degraded environments such as turbidity and partial occlusion and also provides the object range in the scene. A camera array captures the underwater objects in the scene and the temporally encoded binary signals transmitted for the purpose of communication. The network is trained using a clear underwater scene without occlusion, whereas test data is collected in turbid water with partial occlusion. Reconstructed 3D data is the input to a You Look Only Once (YOLOv4) neural network for object detection and a convolutional neural network-based bidirectional long short-term memory network (CNN-BiLSTM) is used for temporal optical signal detection. Finally, the transmitted signal is decoded. In our experiments, 3D InIm provides better image reconstruction in a degraded environment over 2D sensing-based methods. Also, reconstructed 3D images segment out the object of interest from occlusions and background which improves the detection accuracy of the network with 3D InIm. To the best of our knowledge, this is the first report that combines deep learning with 3D InIm for simultaneous and integrated underwater object detection and optical signal detection in degraded environments.

Event-based Spiking Neural Networks for Object Detection: Datasets, Architectures, Learning Rules, and Implementation

Event-based Spiking Neural Networks for Object Detection: Datasets, Architectures, Learning Rules, and Implementation

YoloOW: A Spatial Scale Adaptive Real-Time Object Detection Neural Network for Open Water Search and Rescue From UAV Aerial Imagery

YoloOW: A Spatial Scale Adaptive Real-Time Object Detection Neural Network for Open Water Search and Rescue From UAV Aerial Imagery

Field-of-View IoU for Object Detection in 360° Images.

360° cameras have gained popularity over the last few years. In this paper, we propose two fundamental techniques-Field-of-View IoU (FoV-IoU) and 360Augmentation for object detection in 360° images. Although most object detection neural networks designed for perspective images are applicable to 360° images in equirectangular projection (ERP) format, their performance deteriorates owing to the distortion in ERP images. Our method can be readily integrated with existing perspective object detectors and significantly improves the performance. The FoV-IoU computes the intersection-over-union of two Field-of-View bounding boxes in a spherical image which could be used for training, inference, and evaluation while 360Augmentation is a data augmentation technique specific to 360° object detection task which randomly rotates a spherical image and solves the bias due to the sphere-to-plane projection. We conduct extensive experiments on the 360° indoor dataset with different types of perspective object detectors and show the consistent effectiveness of our method.

MONITORING OF REINTRODUCED RARE PLANTS USING UAV DATA

Abstract. Species reintroduction is one of the most important tasks for biodiversity conservation. In the Samara region, the Department of Ecology, Botany, and Nature Protection of Samara University is constantly working to return rare plants to their natural habitat. Field surveys are carried out to monitor the process of plant development in the natural environment. However, the complex terrain and the special conservation status of the territories, where reintroduction is carried out, require the development of methods for monitoring plants that reduce human contact with the natural environment. As such a method, the article proposes using data from unmanned aerial vehicles (UAVs) in conjunction with the detection of reintroduced plant species using object detection neural networks, namely YOLOv3. The proposed method reduces the anthropogenic impact on the territory during the monitoring process and simplifies plant observations.

Automatic vehicle identification from Google Earth satellite images based on single shot deep learning neural networks for object detection

Automatic vehicle identification from Google Earth satellite images based on single shot deep learning neural networks for object detection

IMAGING: a novel automated system for malaria diagnosis by using artificial intelligence tools and a universal low-cost robotized microscope

IntroductionMalaria is one of the most prevalent infectious diseases in sub-Saharan Africa, with 247 million cases reported worldwide in 2021 according to the World Health Organization. Optical microscopy remains the gold standard technique for malaria diagnosis, however, it requires expertise, is time-consuming and difficult to reproduce. Therefore, new diagnostic techniques based on digital image analysis using artificial intelligence tools can improve diagnosis and help automate it.MethodsIn this study, a dataset of 2571 labeled thick blood smear images were created. YOLOv5x, Faster R-CNN, SSD, and RetinaNet object detection neural networks were trained on the same dataset to evaluate their performance in Plasmodium parasite detection. Attention modules were applied and compared with YOLOv5x results. To automate the entire diagnostic process, a prototype of 3D-printed pieces was designed for the robotization of conventional optical microscopy, capable of auto-focusing the sample and tracking the entire slide.ResultsComparative analysis yielded a performance for YOLOv5x on a test set of 92.10% precision, 93.50% recall, 92.79% F-score, and 94.40% mAP0.5 for leukocyte, early and mature Plasmodium trophozoites overall detection. F-score values of each category were 99.0% for leukocytes, 88.6% for early trophozoites and 87.3% for mature trophozoites detection. Attention modules performance show non-significant statistical differences when compared to YOLOv5x original trained model. The predictive models were integrated into a smartphone-computer application for the purpose of image-based diagnostics in the laboratory. The system can perform a fully automated diagnosis by the auto-focus and X-Y movements of the robotized microscope, the CNN models trained for digital image analysis, and the smartphone device. The new prototype would determine whether a Giemsa-stained thick blood smear sample is positive/negative for Plasmodium infection and its parasite levels. The whole system was integrated into the iMAGING smartphone application.ConclusionThe coalescence of the fully-automated system via auto-focus and slide movements and the autonomous detection of Plasmodium parasites in digital images with a smartphone software and AI algorithms confers the prototype the optimal features to join the global effort against malaria, neglected tropical diseases and other infectious diseases.

RoMP-transformer: Rotational bounding box with multi-level feature pyramid transformer for object detection

This study proposes rotational bounding box with a multi-level feature pyramid transformer (RoMP-Transformer)—a fast and accurate one-stage deep neural network for object detection. The proposed RoMP-Transformer exhibits three characteristics. First, a rotational bounding box is utilized to minimize the effect of the background during the construction of feature maps, enhancing the robustness of the RoMP-Transformer. Second, the RoMP-Transformer employs a multi-level feature pyramid transformer by combining a multi-level feature pyramid network with a pyramid vision-transformer, effectively extracting high-quality features and achieving high accuracy. Third, the RoMP-Transformer executes bounding box optimization by minimizing the optimal intersection of union (IoU) loss by considering both the modified SKEW IoU and distance IoU. The modified SKEW IoU significantly accelerates the calculation, and the fused IoU calculation method improves prediction accuracy. Further, Bayesian optimization and weight lightening with half-tensor are performed to optimize the performance of the RoMP-Transformer for real-time applications. Experiments on three image sets—one on power transmission facilities, MSRA-TD500, and DOTA-v1.0—demonstrate that the proposed RoMP-Transformer outperforms other state-of-the-art neural networks in object detection in terms of accuracy, robustness, and calculation speed. Systematic analysis also reveals that the methods utilized by the RoMP-Transformer optimize object detection performance. The proposed architecture is expected to inspire further study of deep neural networks for object detection in real-world applications.

Crop disease recognition based on the YOLOv5 and transfer learning

Crop disease detection is an important factor in agricultural production. Traditional object detection methods can't effectively screen key features, resulting in weak crop disease control in many countries. In recent years, several convolutional neural networks for object detection have been proposed, which makes it possible to apply computer vision to crop disease identification through deep learning. YOLOv5 is an advanced object detection network, which can extract key features and use human visual attention mechanism as a reference. This paper would evaluate the performance of four pre-trained models of YOLOv5 in object detection of crop diseases. And transfer learning was used to train the corresponding dataset. The experiment results have showed that the F1 values of the four models all reached above 0.93, and the Yolov5x got the best result, which achieved 0.963. Furthermore, the detection accuracy of the four models has reached more than 98%. This shows that the YOLOv5 series network models have great application prospects in the identification of crop diseases. In the near future, the object detection model can be applied to various mobile devices, even unmanned aerial vehicles, which would play a significant role in crop disease prevention.

CoSOV1Net: A Cone- and Spatial-Opponent Primary Visual Cortex-Inspired Neural Network for Lightweight Salient Object Detection.

Salient object-detection models attempt to mimic the human visual system's ability to select relevant objects in images. To this end, the development of deep neural networks on high-end computers has recently achieved high performance. However, developing deep neural network models with the same performance for resource-limited vision sensors or mobile devices remains a challenge. In this work, we propose CoSOV1net, a novel lightweight salient object-detection neural network model, inspired by the cone- and spatial-opponent processes of the primary visual cortex (V1), which inextricably link color and shape in human color perception. Our proposed model is trained from scratch, without using backbones from image classification or other tasks. Experiments on the most widely used and challenging datasets for salient object detection show that CoSOV1Net achieves competitive performance (i.e., Fβ=0.931 on the ECSSD dataset) with state-of-the-art salient object-detection models while having a low number of parameters (1.14 M), low FLOPS (1.4 G) and high FPS (211.2) on GPU (Nvidia GeForce RTX 3090 Ti) compared to the state of the art in lightweight or nonlightweight salient object-detection tasks. Thus, CoSOV1net has turned out to be a lightweight salient object-detection model that can be adapted to mobile environments and resource-constrained devices.

A lightweight neural network for lung nodule detection based on improved ghost module.

Computer tomography images are the preferred method of preoperative evaluation for lung disease. However, it remains difficult to detect and recognize nodules accurately and efficiently due to poor data imaging quality, heavy reliance on physician experience and the need for more human-computer interaction for diagnosis. Currently, image nodule detection based on deep convolutional neural networks has gained much momentum. To alleviate doctors' tremendous labor in the diagnosis procedure, and improve the accuracy of intelligent detection of lung nodules, we improved GhostNet and proposed a lightweight neural network for object detection for lung nodule image detection. Firstly, the bneck structure in the backbone feature extraction network is adopted and improved from the structure of MobileNetV3. The weights are adjusted by changing the initial channel attention mechanism and introducing a spatial-temporal attention mechanism. Then, in the enhanced feature extraction part, we mainly use depth-separable convolution blocks to replace the 3×3 convolution of the original network for the purpose of reducing the model parameters, and make more improvements based on the network structure to enhance the applicability of the network. Diagnostic precision, recall, F1-score, mAP and parameter count were calculated. According to our lightweight neural network, F1-score, precision, and recall were 0.87, 86.34%, and 86.69%, respectively. Based on our dataset, the Yolov4-GNet network proposed in this research outperforms the current neural networks on both precision and recall as well as F1. The lung nodule detection method proposed in this research not only simplifies the processing of images, but also outperforms comparable methods in nodule detection rate and positioning accuracy, providing a new way for lung nodule detection.

Multi-Object Tracking in Heterogeneous environments (MOTHe) for animal video recordings.

Aerial imagery and video recordings of animals are used for many areas of research such as animal behaviour, behavioural neuroscience and field biology. Many automated methods are being developed to extract data from such high-resolution videos. Most of the available tools are developed for videos taken under idealised laboratory conditions. Therefore, the task of animal detection and tracking for videos taken in natural settings remains challenging due to heterogeneous environments. Methods that are useful for field conditions are often difficult to implement and thus remain inaccessible to empirical researchers. To address this gap, we present an open-source package called Multi-Object Tracking in Heterogeneous environments (MOTHe), a Python-based application that uses a basic convolutional neural network for object detection. MOTHe offers a graphical interface to automate the various steps related to animal tracking such as training data generation, animal detection in complex backgrounds and visually tracking animals in the videos. Users can also generate training data and train a new model which can be used for object detection tasks for a completely new dataset. MOTHe doesn't require any sophisticated infrastructure and can be run on basic desktop computing units. We demonstrate MOTHe on six video clips in varying background conditions. These videos are from two species in their natural habitat-wasp colonies on their nests (up to 12 individuals per colony) and antelope herds in four different habitats (up to 156 individuals in a herd). Using MOTHe, we are able to detect and track individuals in all these videos. MOTHe is available as an open-source GitHub repository with a detailed user guide and demonstrations at: https://github.com/tee-lab/MOTHe-GUI.

Deep learning morphological distribution analysis of metal alloy catalysts in proton exchange membrane fuel cells

Measuring the variation of the morphological distribution of Pt-based catalyst particles on supports using transmission electron microscopy provides crucial information for understanding the performance degradation behaviors of proton exchange membrane fuel cells and for designing more durable electrocatalysts. However, interpretation based on a few micrographs is statistically insignificant, whereas manual analyses of large image datasets are time-consuming and often require subjective human decisions. To address this issue, an efficient method assisted by deep learning is proposed for the automated interpretation of massive image datasets of metal catalyst nanoparticles (NPs). Based on an attention-aided deep convolutional neural network for object detection (that is, Attention U-Net), the proposed approach rapidly measures the changes in the structural parameters of Pt/Co NPs on a porous support and quantitatively evaluates morphological changes of the NPs after cycling, with statistical significance for their sizes and separating distances.

Ball Detection Using Deep Learning Implemented on an Educational Robot Based on Raspberry Pi.

RoboCupJunior is a project-oriented competition for primary and secondary school students that promotes robotics, computer science and programing. Through real life scenarios, students are encouraged to engage in robotics in order to help people. One of the popular categories is Rescue Line, in which an autonomous robot has to find and rescue victims. The victim is in the shape of a silver ball that reflects light and is electrically conductive. The robot should find the victim and place it in the evacuation zone. Teams mostly detect victims (balls) using random walk or distant sensors. In this preliminary study, we explored the possibility of using a camera, Hough transform (HT) and deep learning methods for finding and locating balls with the educational mobile robot Fischertechnik with Raspberry Pi (RPi). We trained, tested and validated the performance of different algorithms (convolutional neural networks for object detection and U-NET architecture for sematic segmentation) on a handmade dataset made of images of balls in different light conditions and surroundings. RESNET50 was the most accurate, and MOBILENET_V3_LARGE_320 was the fastest object detection method, while EFFICIENTNET-B0 proved to be the most accurate, and MOBILENET_V2 was the fastest semantic segmentation method on the RPi. HT was by far the fastest method, but produced significantly worse results. These methods were then implemented on a robot and tested in a simplified environment (one silver ball with white surroundings and different light conditions) where HT had the best ratio of speed and accuracy (4.71 s, DICE 0.7989, IoU 0.6651). The results show that microcomputers without GPUs are still too weak for complicated deep learning algorithms in real-time situations, although these algorithms show much higher accuracy in complicated environment situations.

3D Street Object Detection from Monocular Images Using Deep Learning and Depth Information

In this study, we present a three-dimensional (3D) object detection algorithm based on monocular images by constructing an end-to-end network, that incorporates depth information. The entire network consists of three parts. The first part includes the basic object detection neural network as the main body, that uses the region proposal network to obtain the two-dimensional (2D) region proposal of the object. The second part is the depth estimation branch network, that obtains the depth information of the object pixels and calculates the corresponding 3D point cloud. In the last part, concatenated features obtained from the aforementioned two parts are fed into the fully-connected layers. Subsequently, 2D and 3D detection results are obtained. Compared with certain existing methods, the accuracy of the detection results is improved in this study.

Automated visual inspection system for ATLAS ITk strip hybrids

A method of automating the visual inspection of ATLAS upgrade strip modules is shown. The visual inspection of the hybrids is a time consuming part of the quality control during module production. A method of detecting and classifying the SMD components on the hybrids using an object detection neural network was investigated. The results show that the amount of hybrids that needed to be check by a human operator was reduced to around 10% of the batch. This hugely reduced the amount of time needed for human inspection and did find real mistakes done during the production of the hybrids.

Convolutional Neural Networks for Object Detection and Recognition

One of the essential technologies in the fields of target extraction, pattern recognition, and motion measurement is moving object detection. Finding moving objects or a number of moving objects across a series of frames is called object tracking. Basically, object tracking is a difficult task. Unexpected changes in the surroundings, an item's mobility, noise, etc., might make it difficult to follow an object. Different tracking methods have been developed to solve these issues. This paper discusses a number of object tracking and detection approaches. The major methods for identifying objects in images will be discussed in this paper. Recent years have seen impressive advancements in fields like pattern recognition and machine learning, both of which use convolutional neural networks (CNNs). It is mostly caused by graphics processing units'(GPUs) enhanced parallel processing capacity. This article describes many kinds of object classification, object racking, and object detection techniques. Our results showed that the suggested algorithm can detect moving objects reliably and efficiently in a variety of situations.