Enhanced rock recognition via EVSS-integrated YOLO11: A deep learning approach for precise geological classification.

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] With the rapid development of deep learning in computer vision, especially deep convolutional neural networks (CNNs), significant advances have been made in recent years on object recognition and detection in images. Highly accurate detection results have been achieved for large objects, whereas detection accuracy on small objects remains to be low. This dissertation focuses on investigating deep learning methods for small object detection in images and proposing new methods with improved performance. First, we conducted a comprehensive review of existing deep learning methods for small object detections, in which we summarized and categorized major techniques and models, identified major challenges, and listed some future research directions. Existing techniques were categorized into using contextual information, combining multiple feature maps, creating sufficient positive examples, and balancing foreground and background examples. Methods developed in four related areas, generic object detection, face detection, object detection in aerial imagery, and segmentation, were summarized and compared. In addition, the performances of several leading deep learning methods for small object detection, including YOLOv3, Faster R-CNN, and SSD, were evaluated based on three large benchmark image datasets of small objects. Experimental results showed that Faster R-CNN performed the best, while YOLOv3 was a close second. Furthermore, a new deep learning method, called Retina-context Net, was proposed and outperformed state-of-the art one-stage deep learning models, including SSD, YOLOv3 and RetinaNet, on the COCO and SUN benchmark datasets. Secondly, we created a new dataset for bird detection, called Little Birds in Aerial Imagery (LBAI), from real-life aerial imagery. LBAI contains birds with sizes ranging from 10 by 10 pixels to 40 by 40 pixels. We adapted and applied several state-of-the-art deep learning models to LBAI, including object detection models such as YOLOv2, SSH, and Tiny Face, and instance segmentation models such as U-Net and Mask R-CNN. Our empirical results illustrated the strength and weakness of these methods, showing that SSH performed the best for easy cases, whereas Tiny Face performed the best for hard cases with cluttered backgrounds. Among small instance segmentation methods, U-Net achieved slightly better performance than Mask R-CNN. Thirdly, we proposed a new graph neural network-based object detection algorithm, called GODM, to take the spatial information of candidate objects into consideration in small object detection. Instead of detecting small objects independently as the existing deep learning methods do, GODM treats the candidate bounding boxes generated by existing object detectors as nodes and creates edges based on the spatial or semantic relationship between the candidate bounding boxes. GODM contains four major components: node feature generation, graph generation, node class labelling, and graph convolutional neural network model. Several graph generation methods were proposed. Experimental results on the LBDA dataset show that GODM outperformed existing state-of-the-art object detector Faster R-CNN significantly, up to 12% better in accuracy. Finally, we proposed a new computer vision-based grass analysis using machine learning. To deal with the variation of lighting condition, a two-stage segmentation strategy is proposed for grass coverage computation based on a blackboard background. On a real world dataset we collected from natural environments, the proposed method was robust to varying environments, lighting, and colors. For grass detection and coverage computation, the error rate was just 3%.

Deep learning has made significant progress in the field of object detection, especially convolutional neural networks have performed well in image classification, object detection, and segmentation tasks. However, with the increasing complexity of models and the demand for computing resources, traditional deep learning models face challenges in the deployment of resource-constrained mobile and embedded devices. In order to solve this problem, model compression and acceleration techniques have become a research hotspot, including pruning, quantification and knowledge distillation. The purpose of this paper is to review various algorithms in the field of object detection and their advantages and disadvantages, and to discuss the best optimization scheme based on the application and optimization effect of lightweight technology in various algorithms. The research objectives include: systematically summarizing and analyzing the main lightweight technologies currently used for object detection algorithms, evaluating their practical effects in object detection tasks, proposing improvement schemes suitable for specific application scenarios, and looking forward to the future development direction, and discussing potential research directions and technological breakthroughs.

The neural network‐based technologies have emerged as a potent method for image fusion, object detection, and other computer vision tasks as the rapid development of deep learning. Multi‐band infrared images, in particular, capture a more extensive range of radiation details and information compared to conventional single‐band infrared images. Consequently, the fusion of multi‐band infrared images can provide more features for object detection. However, it is crucial to consider that infrared images may contain sensitive information, potentially leading to privacy concerns. Ensuring datasets privacy protection plays a crucial role in the fusion and tracking process. To address both the need for improved detection performance and the necessity for privacy protection in the infrared environment, we proposed a procedure for object detection based on multi‐band infrared image datasets and utilized the transfer learning technique to migrate knowledge learned from external infrared data to internal infrared data, thereby training the infrared image fusion model and detection model. The procedure consists of several steps: (1) data preprocessing of multi‐band infrared images, (2) multi‐band infrared image fusion, and (3) object detection. Standard evaluation metrics for image fusion and object detection ensure the authenticity of the experiments. The comprehensive validation experiments demonstrate the effectiveness of the proposed procedure in object detection tasks. Furthermore, the transfer learning can train our datasets and update the model without exposing the original data. This aspect of transfer learning is particularly beneficial for maintaining the privacy of multi‐band infrared images during the fusion and detection processes.

Underwater object detection is one of the most challenging research topics in computer vision technology. The complex underwater environment makes underwater images suffer from high noise, low visibility, blurred edges, low contrast and color deviation, which brings significant challenges to underwater object detection tasks. In underwater object detection tasks, traditional object detection methods often perform poorly in terms of accuracy and generalization capabilities. Underwater object detection requires accurate, stable, generalizable, real-time and lightweight detection models, for which many researchers have proposed various underwater object detection techniques based on deep learning. Although many outstanding results have been achieved on underwater object detection over the years, the research status of underwater object detection techniques are still lack of unified induction, and some existing problems need to be further probed from the latest perspective. In addition, previous reviews lack analysis on the relationship between underwater image enhancement and object detection. Therefore, this paper provides a comprehensive review of the current research challenges, future development trends, and potential applications of underwater object detection techniques. More importantly, this paper has explored the internal relationship between underwater image enhancement and object detection, and analyzed the possible implementation manners of underwater image enhancement in the object detection task in order to further enhance its benefits. The experiments show the performances of current underwater image enhancement and state-of-the-art object detection algorithms, point out their limitations, and indicate that there is not a strict positive correlation between underwater image enhancement and the accuracy improvement of object detection. The domain shift caused by underwater image enhancement cannot be ignored. This paper can be regarded as a guide for future works on underwater object detection.

FAIR1M: A benchmark dataset for fine-grained object recognition in high-resolution remote sensing imagery

To detect the objects around an autonomous vehicle is very essential to operate safely. This paper presents to detect and classify the objects for assisting autonomous driving. In autonomous driving systems, the task of object detection itself is one of the most important prerequisites to autonomous navigation. Deep learning one of the computer vision tasks, perform object detection very effectively than compared to earlier methods and this project is to detect the objects like vehicles, persons, traffic lights, etc. In this work, an approach to object detection in deep learning that makes the bounding box for an image to predict is explored. Object detection is the method of detecting the objects present in a given image. Apart from detecting the number of objects present in an image it also specifies in which location that object is present in the image. The objects are detected by means of bounding boxes. In the existing system algorithm like Convolutional Neural Network (CNN) using Resnet-50 were used to detect the objects like vehicles, persons, traffic lights separately. The problem identified here is, in the existing system the camera is fixed in a particular place and it detects objects only if the objects come into the camera frames. It is not detecting both objects and lanes simultaneously when the autonomous vehicle is in motion at any location. To overcome these problems, in the proposed system object detection is performed by mounting camera in front of the moving vehicle. You Only Look Once (YOLO) V3 Algorithm is used for the process of object detection. Compared to earlier detection approaches YOLO V3 shows improvement in detection accuracy. It provides good feature extraction and detection in large-scale. The proposed YOLO al gorithm has better average precision value for detecting all objects than compared to existing CNN using Resnet-50. In addition YOLO V3 algorithm performs lane detection apart from object detection.

PurposeData quality and data resolution are essential for computer vision tasks like medical image processing, object detection, pattern recognition and so on. Super-resolution is a way to increase the image resolution, and super-resolved images contain more information compared to their low-resolution counterparts. The purpose of this study is analyzing the effects of the super resolution models trained before on object detection for aerial images.Design/methodology/approachTwo different models were trained using the Super-Resolution Generative Adversarial Network (SRGAN) architecture on two aerial image data sets, the xView and the Dataset for Object deTection in Aerial images (DOTA). This study uses these models to increase the resolution of aerial images for improving object detection performance. This study analyzes the effects of the model with the best perceptual index (PI) and the model with the best RMSE on object detection in detail.FindingsSuper-resolution increases the object detection quality as expected. But, the super-resolution model with better perceptual quality achieves lower mean average precision results compared to the model with better RMSE. It means that the model with a better PI is more meaningful to human perception but less meaningful to computer vision.Originality/valueThe contributions of the authors to the literature are threefold. First, they do a wide analysis of SRGAN results for aerial image super-resolution on the task of object detection. Second, they compare super-resolution models with best PI and best RMSE to showcase the differences on object detection performance as a downstream task first time in the literature. Finally, they use a transfer learning approach for super-resolution to improve the performance of object detection.

Abstract. At the time of continuous development of all technologies, deep machine learning (more precisely, convolutional neural networks), which is one of the branches of artificial intelligence (AI), has found wide application in many fields, including photogrammetry and remote sensing. One of the areas where a lot of research is conducted using these methods is the recognition of objects in aerial and satellite imagery. Through the application of deep learning algorithms and neural networks, it is possible to automate labour-intensive processes. However, while object detection in images using machine learning is popular for natural scenes and in recent years also for nadir aerial and satellite imagery, for aerial oblique imagery at the moment of this research there were relatively few publications on the subject. The challengeable task in object detection is the time-consuming generation of training datasets when access is limited or non-existent. This study proposed the methodology to automate this process with use of existing resources for transferring of references to new databases for training models for detect objects on aerial oblique images. The object detection was performed using the YOLOv3 neural network. Experiment results tested on two datasets have shown that the proposed method could realize the task of object detection in oblique aerial images.

Learning to capture dependencies between global features of different convolution layers

General deep learning methods achieve high-level semantic feature representation by aggregating hierarchical features, which performs well in object detection tasks. However, issues arise with general deep learning methods in UAV-based remote sensing image object detection tasks. Firstly, general feature aggregation methods such as stride convolution may lead to information loss in input samples. Secondly, common FPN methods introduce conflicting information by directly fusing feature maps from different levels. These shortcomings limit the model’s detection performance on small and weak targets in remote sensing images. In response to these concerns, we propose an unmanned aerial vehicle (UAV) object detection algorithm, IF-YOLO. Specifically, our algorithm leverages the Information-Preserving Feature Aggregation (IPFA) module to construct semantic feature representations while preserving the intrinsic features of small objects. Furthermore, to filter out irrelevant information introduced by direct fusion, we introduce the Conflict Information Suppression Feature Fusion Module (CSFM) to improve the feature fusion approach. Additionally, the Fine-Grained Aggregation Feature Pyramid Network (FGAFPN) facilitates interaction between feature maps at different levels, reducing the generation of conflicting information during multi-scale feature fusion. The experimental results on the VisDrone2019 dataset demonstrate that in contrast to the standard YOLOv8-s, our enhanced algorithm achieves a mean average precision (mAP) of 47.3%, with precision and recall rates enhanced by 6.3% and 5.6%, respectively.

计算机视觉的任务目标是建立接近人类视觉系统的计算模型。随着深度神经网络（deep neural network，DNN）的发展，对计算机视觉中高层语义的分析与理解成为研究重点。计算机视觉的高层语义通常为人类可理解、可表述的用于表达图像、视频等媒体信号内容的描述子（descriptor），典型的高层语义分析任务包含图像分类、目标检测、实例分割、语义分割与视频场景识别、目标跟踪等。基于深度神经网络的算法使计算机视觉任务获得逐步提升的性能，但是网络模型的体量增大与计算效率的降低随之而来。模型蒸馏是一种基于迁移学习进行模型压缩的方案。此类方案通常利用一个预训练模型作为教师，提取其有效的表示，如模型输出、隐藏层特征或特征间相似度等，并将上述表示作为另一个规模较小、推断速度较快的学生模型的额外监督信号，对该学生模型进行训练，以达到提升小模型性能从而取代大模型的目的。模型蒸馏对模型性能与计算复杂度有着良好权衡，因此愈来愈多地用于基于深度学习的高层语义分析中。自2014年模型蒸馏概念提出以来，研究人员开发了大量应用于高层语义分析的模型蒸馏方法，在图像分类、目标检测与语义分割任务中的应用最为广泛。本文对上述典型任务中具有代表性的模型蒸馏方案进行调研和汇总，依照不同的视觉任务进行介绍。首先，从最成熟、应用最广泛的分类任务模型蒸馏方法开始，介绍其不同的设计思路与应用场景，展示部分实验性能的对比，指出在分类任务上与在检测、分割任务上应用模型蒸馏的条件差异性。接着，对几种经特殊设计而应用于目标检测、语义分割的典型模型蒸馏方法进行介绍，结合模型结构对设计目的与思路进行说明，提供部分实验结果的对比与分析。最后，对当前高层语义分析中模型蒸馏方法的现状进行了总结分析，并指出存在的困难及不足，设想未来可能的探索思路与发展方向。

The detection of transparent object such as glass in the image is recently popular in computer vision researches. Among the various tasks of detecting objects in images, it is not an easy task to detect the presence of transparent objects in the image. The detection of transparent objects is very difficult to perform using classical computer vision algorithms since the appearance of transparent objects dramatically depends on its background and illumination conditions. In addition to the popularity of transparent object detection, deep learning is also giving high performance in object detection tasks. In this paper, we apply one of the Convolutional Neural Network called Single Shot MultiBox Detector (SSD) for transparent object detection task and evaluate the performance of the system. The results show that the application of deep learning method in detection of transparent objects can successfully perform the detection of transparent objects in images.

A 3D-CAE-CNN model for Deep Representation Learning of 3D images

Object detection is a hot topic in the field of computer vision and pattern recognition. The task of object detection is to accurately and efficiently identify and locate many object instances of predefined categories from images. With the wide application of deep learning, the accuracy and efficiency of object detection have been greatly improved. However, object detection based on deep learning still faces challenges such as improving the performance of mainstream object detection algorithms and the detection accuracy of small target objects. In this paper, based on extensive literature research, we survey the mainstream algorithms of object detection from the angle of improving and optimizing the two-stage and onestage object detection algorithms. We also analyze the promotion method of small object detection accuracy combined with the backbone network, the visual receptive field, and the model's training. In addition, the common data sets of object detection are introduced in detail, while the performance of representative algorithms is compared from two aspects. The problems to be solved in object detection and the future research direction are predicted and prospected. More high precision and efficient algorithms are proposed, and more research directions will be developed in the future.

Aiming at the problems of limited feature expression and underutilization of contextual information in the object detection task in low-light environments, a low-light object detection algorithm based on feature complementarity and multicontent aggregation (FCMA-Det) is proposed in this article. The algorithm is based on the FCOS framework and improves the model performance through a stepwise optimization strategy. First, the feature complementary module (FCM) is designed. By dynamically fusing the dual-domain features of the original low-light image and the augmented image, the feature discriminative property of the augmented domain is strengthened while retaining the detail information of the original domain, so as to solve the problem of insufficient single-domain feature expression capability. Second, a multicontent aggregation module (MAM) is constructed. Combining multiscale context fusion and adaptive sensory field adjustment mechanism, it accurately captures the local details of the object and long-range context dependencies, and effectively copes with the challenges of complex background interference and insufficient feature extraction. In addition, feature consistency loss and dynamic weight fusion loss function are introduced to further optimize cross-domain feature alignment and multibranch collaborative training. Experiments on low-light dataset such as Exclusively Dark (ExDark) and small-object mixed-light dataset such as tiny object detection in aerial images (AI-TOD) show that the mean average precision (mAP) metric of FCMA-Det is improved by 1.7% on average compared with the existing SOTA method, which verifies the robustness and generalization ability of the algorithm under extreme light conditions.