Faster R-CNN and YOLOv3: a general analysis between popular object detection networks

Algorithms based on deep learning are taking off as solutions to several problems in numerous fields. Numerous deep learning architectures like convolutional neural networks, recurrent neural networks, deep belief networks, artificial neural networks and deep neural networks have been used in fields such as audio recognition, machine vision, processing of natural language, recognition of speech, social network filtering, computer vision, translating languages and bioinformatics. One such application of deep learning architectures is object detection. Object detection in satellite and aerial imagery undertakes an important role in a long list of applications and has allured a lot of attention which resulted in exceptional progress in recent years. It has abundant uses in various fields such as surveillance, military, national security, urban planning, environmental applications, traffic control and forestry. Several issues like varying size, background and orientation of the target object make the automatic detection of target in satellite imagery a challenging issue. Recent research in this field has given better speed and accuracy to the detection of the target object. With the rise in the data being generated by the Earth Observation programs, it is becoming much more imperative to find better and efficient ways to exploit this data by improving upon the existing Deep Learning research. In this paper, we bring forth a wide-ranging analysis of the recent advancements in methods of object detection based on deep learning to recognize the objects in satellite imagery.KeywordsDeep learningObject detectionSatellite imageryRemote sensingYou Only Look Once (YOLO)You Only Look Twice (YOLT)Single Shot Detector (SSD)Convolutional Neural Networks (CNN)Faster Region-based Convolutional Neural Networks (Faster RCNN)Hybrid Deep Convolutional Neural Network (HDNN)Satellite Imagery Multiscale Rapid Detection with Windowed Networks (SIMRDWN)

In the last couple of years, advancements in the deep learning, especially in convolutional neural networks, proved to be a boon for the image classification and recognition tasks. One of the important practical applications of object detection and image classification can be for security enhancement. If dangerous objects or scenes can be identified automatically, then a lot of accidents can be prevented. For this purpose, in this paper we made use of state-of-the-art implementation of Faster Region-based Convolutional Neural Network (Faster R-CNN) based on the monitoring video of hoisting sites to train a model to detect the dangerous object and the worker. By extracting the locations of them, object-human interactions during hoisting, mainly for changes in their spatial location relationship, can be understood whereby estimating whether the scene is safe or dangerous. Experimental results showed that the pre-trained model achieved good performance with a high mean average precision of 97.66% on object detection and the proposed method fulfilled the goal of dangerous scenes recognition perfectly.

<title>Abstract</title> Accurate and efficient object detection is essential for intelligent road-scene monitoring systems operating in visually complex and resource-constrained environments. While one-stage detectors achieve high inference speed, they often struggle with precise localization of small or low-contrast objects, whereas two-stage detectors provide higher accuracy at the cost of increased latency. To address this trade-off, this paper proposes a hybrid object-detection architecture that integrates You Only Look Once version 5-Small (YOLOv5s) as a fast proposal generator with Faster Region-Based Convolutional Neural Network (Faster R-CNN) as a region-wise refinement module. The proposed framework replaces the Region Proposal Network of Faster R-CNN with high-confidence YOLOv5s detections and employs confidence-weighted fusion to produce spatially consistent final predictions. The hybrid model was evaluated on complex road-scene data using standard object-detection metrics, including mean Average Precision at IoU 0.50 (mAP@50), precision, recall, and inference speed. Experimental results show that the proposed approach achieves mAP@50 of 0.89, improving upon the YOLOv5s baseline by 4.7 percentage points, while maintaining near–real-time performance at 45 frames per second, which is approximately three times faster than a standalone Faster R-CNN. The hybrid detector also attained a precision of 0.93 and a recall of 0.90, demonstrating improved localization accuracy and reduced false detections, particularly for small and visually ambiguous road-scene objects. Repeated experiments confirmed the robustness of the approach, with consistent accuracy gains and low variance across runs. These results demonstrate that strategically combining one-stage and two-stage detection paradigms can yield a favorable accuracy–efficiency balance, making the proposed hybrid architecture suitable for practical deployment in intelligent road-infrastructure and smart-city applications.

Detection of metastasis of mediastinal lymph nodes in lung cancer patients with an artificial intelligence model.

In the past few years, automatic building detection in aerial images has become an emerging field in computer vision. Detecting the specific types of houses will provide information in urbanization, change detection, and urban monitoring that play increasingly important roles in modern city planning and natural hazard preparedness. In this paper, we demonstrate the effectiveness of detecting various types of houses in aerial imagery using Faster Region-based Convolutional Neural Network (Faster-RCNN). After formulating the dataset and extracting bounding-box information, pre-trained ResNet50 is used to get the feature maps. The fully convolutional Region Proposal Network (RPN) first predicts the bounds and objectness score of objects (in this case house) from the feature maps. Then, the Region of Interest (RoI) pooling layer extracts interested regions to detect objects that are present in the images. To the best of our knowledge, this is the first attempt at detecting houses using Faster R-CNN that has achieved satisfactory results. This experiment opens a new path to conduct and extent the works not only for civil and environmental domain but also other applied science disciplines.

This paper describes the use of two different deep-learning algorithms for object detection to recognize different badgers. We use recordings of four different badgers under varying background illuminations. In total four different object detection algorithms based on deep neural networks are compared: The single shot multi-box detector (SSD) with the Inception-V2 or MobileNet as a backbone, and the faster region-based convolutional neural network (Faster R-CNN) combined with Inception-V2 or residual networks. Furthermore, two different activation functions are compared to compute probabilities that some badger is in the detected region: the softmax and sigmoid functions. The results of all eight models show that SSD obtains higher recognition accuracies (97.8%–98.6%) than Faster R-CNN (84.8%–91.7%). However, the training time of Faster R-CNN is much shorter than that of SSD. The use of different output activation functions seems not to matter much.

Complex background classification network: A deep learning method for urban images classification

The aim of this study was to compare the performance of a deep-learning convolutional neural network (Faster R-CNN) model to detect imaging findings suggestive of idiopathic Parkinson’s disease (PD) based on [18F]FP-CIT PET maximum intensity projection (MIP) images versus that of nuclear medicine (NM) physicians. The anteroposterior MIP images of the [18F]FP-CIT PET scan of 527 patients were classified as having PD (139 images) or non-PD (388 images) patterns according to the final diagnosis. Non-PD patterns were classified as overall-normal (ONL, 365 images) and vascular parkinsonism with definite defects or prominently decreased dopamine transporter binding (dVP, 23 images) patterns. Faster R-CNN was trained on 120 PD, 320 ONL, and 16 dVP pattern images and tested on the 19 PD, 45 ONL, and seven dVP patterns images. The performance of the Faster R-CNN and three NM physicians was assessed using receiver operating characteristics curve analysis. The difference in performance was assessed using Cochran’s Q test, and the inter-rater reliability was calculated. Faster R-CNN showed high accuracy in differentiating PD from non-PD patterns and also from dVP patterns, with results comparable to those of NM physicians. There were no significant differences in the area under the curve and performance. The inter-rater reliability among Faster R-CNN and NM physicians showed substantial to almost perfect agreement. The deep-learning model accurately differentiated PD from non-PD patterns on MIP images of [18F]FP-CIT PET, and its performance was comparable to that of NM physicians.

Foreign objects such as packaging bags on the road pose a significant threat to driving safety, especially at high speeds or under low-visibility conditions. However, research on detecting road packaging bags remains limited, and existing object detection models face challenges in small object detection, computational efficiency, and embedded deployment. To address these issues, the lightweight deep learning model RGE-YOLO is the foundation for the real-time detection technique proposed in this contribution. Built upon YOLOv8s, RGE-YOLO incorporates RepViTBlock, Grouped Spatial Convolution (GSConv), and Efficient Multi-Scale Attention (EMA) to optimize computational efficiency, model stability, and detection accuracy. GSConv reduces redundant computations, enhancing model lightweight; EMA enhances the model’s ability to capture multi-scale information by integrating channel and spatial attention mechanisms; RepViTBlock integrates convolution and self-attention mechanisms to improve feature extraction capabilities. The proposed method was validated on a custom-built road plastic bag dataset comprising 6,000 augmented images. Experimental results demonstrate that RGE-YOLO outperforms state-of-the-art models such as Single Shot MultiBox Detector (SSD) and Faster Region-based Convolutional Neural Network (Faster R-CNN) in terms of mean average precision (mAP 92.2%) and detection speed (250 FPS), while significantly reducing model parameters (9.1 M) and computational complexity (23.9 GFLOPs), increasing its suitability for installation on computerized systems within vehicles. It introduces an effective and lightweight approach for detecting road packaging bags and contributes to increased driving safety.

Detection of threat objects concealed in passenger clothing and baggage poses a challenge to aviation security. At present, the detection technology is capable of detecting the presence of threats from the scanned images yet requires the involvement of human in determining what type of threat and where it is located. Deep learning-based object detection technique has not been successfully implemented to detect threats in the security screening processes. This research aims to improve the accuracy and the processing time of threat detection in security screening. Enhanced faster region-based convolutional neural network (faster R-CNN) with improved region proposals is used for better threat localization. The proposed system consists of an improved region proposal network that outputs object’s region proposals with an object score to the detector module to accurately locate the threat in the human body. Furthermore, this system uses a modified loss function that strengthens the classification loss. Results obtained by the proposed model show a 15% improvement in object localization. Therefore, the enhanced faster R-CNN achieves an overall detection accuracy of 0.27 in terms of average precision and reduces the processing time by 0.19 s. The results obtained by the enhanced faster R-CNN for detection accuracy are superior to the state-of-the-art system. Also, this model focuses on localizing the threat and identifying its type, which makes the model suitable for threat detection security screening. Besides, the research also addresses the time consumption issue in detecting the threat object.

Currently, visual inspection techniques, especially closed-circuit television (CCTV), are commonly utilized for sewer pipe inspection. Computer vision techniques are applied for automated interpretation of CCTV images to identify pipe defects. However, conventional computer vision techniques require complex handcrafted feature extraction and large amount of image pre-processing. In this study, a deep learning based approach is developed for sewer pipe defect detection using faster region-based convolutional neural network (faster R-CNN). 3000 images were collected from CCTV inspection videos of sewer pipes, among which 85% were used for training and validation and 15% are for testing. The detection model was trained and evaluated in terms of mean average precision (mAP), missing rate, detection speed and training time. The proposed approach is demonstrated to be applicable for detecting sewer pipe defects accurately with a high mAP and low missing rate. In addition, the initial model was improved by investigating the influence of dataset size, initialization network type and training mode, as well as network hyper-parameters on model performance. The improved model achieved a mAP of 83% and fast detection speed. This study has the potential for addressing similar object detection problems in the architecture, engineering and construction (AEC) industry and provides references when designing the deep learning models.

In recent years, developments in artificial intelligence (AI) and computer vision (CV) have led to an increase in research regarding retail product detection. However, effective CV use for identifying objects continues to face major obstacles, such as customers% random placement of products at checkout partially obscuring items from view. This paper explores the use of faster region-based convolutional neural networks (Faster RCNN) and attentional reweighting to detect obscured retail products. This attention-based method allows CV to identify and process various occlusion patterns. While the training set used to teach the network excluded heavily occluded data, the test set included a range of obscuration levels to accurately reflect the random positioning of products in real life. Experimental results suggest that the proposed method returns better detection accuracy with occlusion problems than current approaches employing convolutional neural networks (CNNs).

Vehicle weight identification system for spatiotemporal load distribution on bridges based on non-contact machine vision technology and deep learning algorithms

This study aims to improve the accuracy of landslide detection in satellite images by combining two object detection models based on a faster region-based convolutional neural network (Faster R-CNN) with a classification decision tree. The proposed method combines the predicted results from the two Faster R-CNN models and classifies their features with a classification decision tree to generate a bounding-box that surrounds the landslide area in the input image. The first Faster R-CNN model is trained by using a training set of color images (RGB images). The second model is trained by using grayscale images that represent digital elevation models (DEMs). The results from both models are used to construct features for training a classification decision tree. The resulting bounding-box is selected from the following four classes: the box obtained from the RGB model, the box obtained from the DEM model, the intersection of those two boxes, and the smallest box that contains the union of them. The evaluation results show that the proposed method is better than the RGB model in terms of accuracy, precision, recall, F-measure, and Intersection-over-Union (IoU) score. It is slightly better than the DEM model in almost all evaluation metrics, except the precision.

Apple leaf diseases seriously affect the sustainable production of apple fruit. Early infection monitoring of apple leaves and timely disease control measures are the key to ensuring the regular growth of apple fruits and achieving a high-efficiency economy. Consequently, disease detection schemes based on computer vision can compensate for the shortcomings of traditional disease detection methods that are inaccurate and time-consuming. Nowadays, to solve the limitations ranging from complex background environments to dense and small characteristics of apple leaf diseases, an improved Faster region-based convolutional neural network (Faster R-CNN) method was proposed. The advanced Res2Net and feature pyramid network architecture were introduced as the feature extraction network for extracting reliable and multi-dimensional features. Furthermore, RoIAlign was also employed to replace RoIPool so that accurate candidate regions will be produced to address the object location. Moreover, soft non-maximum suppression was applied for precise detection performance of apple leaf disease when making inferences to the images. The improved Faster R-CNN structure behaves effectively in the annotated apple leaf disease dataset with an accuracy of 63.1% average precision, which is higher than other object detection methods. The experiments proved that our improved Faster R-CNN method provides a highly precise apple leaf disease recognition method that could be used in real agricultural practice.