Focal Iou loss: More attentive learning for bounding box regression

The regression loss function is a key factor in the training and optimization process of object detection. The current mainstream regression loss functions are Ln norm loss, IOU loss and CIOU loss. This paper proposes the Scale-Sensitive IOU(SIOU) loss, a new loss function different from the above all, which could solve the issues that the current loss functions cannot distinguish the two bounding boxes in some special cases when the target area scales in one image vary greatly during training process, thereby leading to the improper regression loss calculation and the slowing down of the optimization. An area scale regulating factor Y is added on the basis of CIOU loss to adjust the loss values of the bounding boxes, which could distinguish all the boxes quantitatively in theory thus gets a faster converging speed and better optimization. Through analysis and simulation comparison among the several loss functions, the superiority of SIOU loss is verified. Furthermore, by incorporating SIOU loss into YOLO v4, Faster R-CNN and SSD on the two mainstream aerial remote sensing datasets, i.e., DIOR and NWPU VHR-10, the detection precisions improve by 10.2% than IOU loss and 2.8% than CIOU loss respectively.

Aerial image-based target detection has problems such as low accuracy in multiscale target detection situations, slow detection speed, missed targets and falsely detected targets. To solve this problem, this paper proposes a detection algorithm based on the improved You Only Look Once (YOLO)v3 network architecture from the perspective of model efficiency and applies it to multiscale image-based target detection. First, the K-means clustering algorithm is used to cluster an aerial dataset and optimize the anchor frame parameters of the network to improve the effectiveness of target detection. Second, the feature extraction method of the algorithm is improved, and a feature fusion method is used to establish a multiscale (large-, medium-, and small-scale) prediction layer, which mitigates the problem of small target information loss in deep networks and improves the detection accuracy of the algorithm. Finally, label regularization processing is performed on the predicted value, the generalized intersection over union (GIoU) is used as the bounding box regression loss function, and the focal loss function is integrated into the bounding box confidence loss function, which not only improves the target detection accuracy but also effectively reduces the false detection rate and missed target rate of the algorithm. An experimental comparison on the RSOD and NWPU VHR-10 aerial datasets shows that the detection effect of high-efficiency YOLO (HE-YOLO) is significantly improved compared with that of YOLOv3, and the average detection accuracies are increased by 8.92% and 7.79% on the two datasets, respectively. The algorithm not only shows better detection performance for multiscale targets but also reduces the missed target rate and false detection rate and has good robustness and generalizability.

Respiratory diseases constitute one of the leading causes of death worldwide and directly affect the patient’s quality of life. Early diagnosis and patient monitoring, which conventionally include lung auscultation, are essential for the efficient management of respiratory diseases. Manual lung sound interpretation is a subjective and time-consuming process that requires high medical expertise. The capabilities that deep learning offers could be exploited in order that robust lung sound classification models can be designed. In this paper, we propose a novel hybrid neural model that implements the focal loss (FL) function to deal with training data imbalance. Features initially extracted from short-time Fourier transform (STFT) spectrograms via a convolutional neural network (CNN) are given as input to a long short-term memory (LSTM) network that memorizes the temporal dependencies between data and classifies four types of lung sounds, including normal, crackles, wheezes, and both crackles and wheezes. The model was trained and tested on the ICBHI 2017 Respiratory Sound Database and achieved state-of-the-art results using three different data splitting strategies—namely, sensitivity 47.37%, specificity 82.46%, score 64.92% and accuracy 73.69% for the official 60/40 split, sensitivity 52.78%, specificity 84.26%, score 68.52% and accuracy 76.39% using interpatient 10-fold cross validation, and sensitivity 60.29% and accuracy 74.57% using leave-one-out cross validation.

Probability calibration and decision threshold selection are fundamental aspects of risk prediction and classification, respectively. A strictly proper loss function is used in clinical risk prediction applications to encourage a model to predict calibrated class-posterior probabilities or risks. Recent studies have shown that training with focal loss can improve the discriminatory power of gradient-boosted decision trees (GBDT) for classification tasks with an imbalanced or skewed class distribution. However, the focal loss function is not a strictly proper loss function. Therefore, the output of GBDT trained using focal loss is not an accurate estimate of the true class-posterior probability. This study aims to address the issue of poor calibration of GBDT trained using focal loss in the context of clinical risk prediction applications. The methodology utilizes a closed-form transformation of the confidence scores of GBDT trained with focal loss to estimate calibrated risks. The closed-form transformation relates the focal loss minimizer and the true-class posterior probability. Algorithms based on Bayesian hyperparameter optimization are provided to choose the focal loss parameter that optimizes discriminatory power and calibration, as measured by the Brier score metric. We assess how the calibration of the confidence scores affects the selection of a decision threshold to optimize the balanced accuracy, defined as the arithmetic mean of sensitivity and specificity. The effectiveness of the proposed strategy was evaluated using lung transplant data extracted from the Scientific Registry of Transplant Recipients (SRTR) for predicting post-transplant cancer. The proposed strategy was also evaluated using data from the Behavioral Risk Factor Surveillance System (BRFSS) for predicting diabetes status. Probability calibration plots, calibration slope and intercept, and the Brier score show that the approach improves calibration while maintaining the same discriminatory power according to the area under the receiver operating characteristics curve (AUROC) and the H-measure. The calibrated focal-aware XGBoost achieved an AUROC, Brier score, and calibration slope of 0.700, 0.128, and 0.968 for predicting the 10-year cancer risk, respectively. The miscalibrated focal-aware XGBoost achieved equal AUROC but a worse Brier score and calibration slope (0.140 and 1.579). The proposed method compared favorably to the standard XGBoost trained using cross-entropy loss (AUROC of 0.755 versus 0.736 in predicting the 1-year risk of cancer). Comparable performance was observed with other risk prediction models in the diabetes prediction task.

To construct a visual intelligent recognition model for Oncomelania hupensis robertsoni in Yunnan Province based on the EfficientNet-B4 model, and to evaluate the impact of data augmentation methods and model hyperparameters on the recognition of O. hupensis robertsoni. A total of 400 O. hupensis robertsoni and 400 Tricula snails were collected from Yongsheng County, Yunnan Province in June 2024, and snail images were captured following identification and classification of 300 O. hupensis robertsoni and 300 Tricula snails. A total of 925 O. hupensis robertsoni images and 1 062 Tricula snail images were collected as a dataset and divided into a training set and a validation set at a ratio of 8:2, while 352 images captured from the remaining 100 O. hupensis robertsoni and 354 images from the remaining 100 Tricula snails served as an external test set. All acquired images were subjected to preprocessing, including cropping and resizing. Three data augmentation approaches were employed, including baseline, Mixup and Gaussian blurring, and model hyperparameters included two optimization algorithms of adaptive moment estimation (Adam) and stochastic gradient descent (SGD), two loss functions of focal loss and cross entropy loss, and two learning rate decay strategies of cosine annealing and multi-step. The intelligent recognition models of O. hupensis robertsoni and Tricula snails were constructed based on the EfficientNet-B4 model, and 7 training strategy groups were generated by combinations of different data augmentation approaches and hyperparameters. The performance of intelligent recognition models was tested with external test sets, and evaluated with accuracy, precision, recall, F1 score, loss, Youden's index, and the area under the receiver operating characteristic curve (AUC) under different training strategies. The variation of loss values was comparable among intelligent recognition models with different data augmentation approaches. The Group 4 model constructed with Mixup and Gaussian blurring data augmentation approaches showed the optimal performance, with an accuracy of 90.38%, precision of 90.07%, F1 score of 89.44%, Youden's index of 0.81 and AUC of 0.961 in the external test set. The accuracy of models using the SGD optimizer reduced by 29.16% as compared to those using the Adam optimizer (χ2 = 81.325, P < 0.001), and the accuracy of models using the cross entropy loss function reduced by 0.80% as compared to the Group 4 model (χ2 = 3.147, P > 0.05), while the accuracy of models using the multi-step learning rate decay strategy increased by 0.65% as compared to the Group 4 model (χ2 = 0.208, P > 0.05). In addition, the model with the baseline + Mixup + Gaussianblurring data augmentation approach and hyperparameters of Adam optimizer, focal loss function and multi-step learning rate decay strategy showed the highest performance, with an accuracy of 91.03%, precision of 91.97%, recall of 88.11%, F1 score of 90.00%, Youden's index of 0.82 and AUC values of 0.969 in external test set, respectively. The intelligent recognition model of O. hupensis robertsoni based on EfficientNet-B4 model is accurate for identification of O. hupensis robertsoni and Tricula snails in Yunnan Province.

Asymmetric double-winged multi-view clustering network for exploring diverse and consistent information

The importance of network security has become increasingly prominent due to the rapid development of network technology. Network intrusion detection systems (NIDSs) play a crucial role in safeguarding networks from malicious attacks and intrusions. However, the issue of class imbalance in the dataset presents a significant challenge to NIDSs. In order to address this concern, this paper proposes a new NIDS called CBF-IDS, which combines convolutional neural networks (CNNs) and bidirectional long short-term memory networks (BiLSTMs) while employing the focal loss function. By utilizing CBF-IDS, spatial and temporal features can be extracted from network traffic. Moreover, during model training, CBF-IDS applies the focal loss function to give more weight to minority class samples, thereby mitigating the impact of class imbalance on model performance. In order to evaluate the effectiveness of CBF-IDS, experiments were conducted on three benchmark datasets: NSL-KDD, UNSW-NB15, and CIC-IDS2017. The experimental results demonstrate that CBF-IDS outperforms other classification models, achieving superior detection performance.

Accurate segmentation of abdominal organs on MRI is crucial for computer-aided surgery and computer-aided diagnosis. Most state-of-the-art methods for MRI segmentation employ an encoder-decoder structure, with skip connections concatenating shallow features from the encoder and deep features from the decoder. In this work, we noticed that simply concatenating shallow and deep features was insufficient for segmentation due to the feature gap between shallow features and deep features. To mitigate this problem, we quantified the feature gap from spatial and semantic aspects and proposed a spatial loss and a semantic loss to bridge the feature gap. The spatial loss enhanced spatial details in deep features, and the semantic loss introduced semantic information into shallow features. The proposed method successfully aggregated the complementary information between shallow and deep features by formulating and bridging the feature gap. Experiments on two abdominal MRI datasets demonstrated the effectiveness of the proposed method, which improved the segmentation performance over a baseline with nearly zero additional parameters. Particularly, the proposed method has advantages for segmenting organs with blurred boundaries or in a small scale, achieving superior performance than state-of-the-art methods.

Compared with natural images, remote sensing targets have small and dense target shapes as well as complex target backgrounds. As a result, insufficient detection accuracy and target location cannot be accurately identified. So, this paper proposes the YOLO-extract algorithm based on the YOLOv5 algorithm. Firstly, The YOLO-extract algorithm optimized the model structure of the YOLOv5 algorithm. The YOLO-extract algorithm not only deleted the feature layer and prediction head with poor feature extraction ability but also a new feature extractor with stronger feature extraction ability was integrated into the network. At the same time, YOLO-extract borrowed the idea of residual network to integrate Coordinate Attention into the network. Secondly, the mixed dilated convolution was combined with the redesigned residual structure to enhance the feature and location information extraction ability of the shallow layer of the model and optimize the feature extraction ability of the model for different scale targets. Finally, drawing on the idea of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> -IoU Loss, Focal- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\alpha $ </tex-math></inline-formula> EIoU Loss was designed to replace CIoU Loss, which makes the model bounding box regression faster and the loss lower. The experimental results on the test data set show that compared with the YOLOv5 algorithm, the YOLO-extract algorithm has a faster convergence speed, reduces the calculation amount by 45.3GFLOPs and the number of parameters by 10.526M, but increases the mAP by 8.1% and the detection speed by 3 times.

In recent years, in the field of remote sensing image target detection, oriented target detection has received extensive attention. Traditional detectors use horizontal bounding box (HBB) to locate targets, which may cause inaccurate detection when detecting targets with arbitrary orientation angles, dense distribution and large aspect ratios. The Oriented Bounding Box (OBB) adds different rotation angles to the horizontal bounding box, which can better deal with the above problems. However, with the introduction of oriented bounding boxes in the detector, new problems have emerged, such as the increase in the number of anchor points with different scales, angles, and aspect ratios, and the imbalance of IoU matching of positive and negative anchor boxes, which leads to classification confidence and Inconsistency between positioning accuracy. In order to overcome these shortcomings while using the oriented bounding box, this paper proposes a novel oriented target detector and redesigns a matching strategy for evaluating the position of the anchor frame. By selecting a higher quality anchor frame to achieve more accurate target detection, thereby reducing the instability of the angle in the matching process. Experiments conducted on public remote sensing data sets such as HRSC2016 and UCAS-AOD show that the proposed method achieves advanced detection accuracy.

Island building change detection is a critical technology for environmental monitoring, disaster early warning, and urban planning, playing a key role in dynamic resource management and sustainable development of islands. However, the imbalanced distribution of class pixels (changed vs. unchanged) undermines the detection capability of existing methods and severe boundary misdetection. To address issue, we propose the MSDT-Net model, which makes breakthroughs in architecture, modules, and loss functions; a dual-branch twin ConvNeXt architecture is adopted as the feature extraction backbone, and the designed Edge-Aware Smoothing Module (MSA) effectively enhances the continuity of the change region boundaries through a multi-scale feature fusion mechanism. The proposed Difference Feature Enhancement Module (DTEM) enables deep interaction and fusion between original semantic and change features, significantly improving the discriminative power of the features. Additionally, a Focal–Dice–IoU Boundary Joint Loss Function (FDUB-Loss) is constructed to suppress massive background interference using Focal Loss, enhance pixel-level segmentation accuracy with Dice Loss, and optimize object localization with IoU Loss. Experiments show that on a self-constructed island dataset, the model achieves an F1-score of 0.9248 and an IoU value of 0.8614. Compared to mainstream methods, MSDT-Net demonstrates significant improvements in key metrics across various aspects. Especially in scenarios with few changed pixels, the recall rate is 0.9178 and the precision is 0.9328, showing excellent detection performance and boundary integrity. The introduction of MSDT-Net provides a highly reliable technical pathway for island development monitoring.

During the drone aerial photography process, due to factors such as diverse target scales and dense occlusion, lead to frequent missed and false detection of small targets. We noticed that the feature fusion module in the YOLOv7 network only simply concatenates channel dimensions for features at different levels, disregarding the influence of feature significance at varying levels on the amalgamation procedure. To tackle the aforementioned concerns, we suggest a straightforward and efficient enhanced framework - the multiscale feature dynamic fusion network MFDF-YOLOv7 based on YOLOv7. Firstly, a multi-scale feature dynamic fusion module (MFDF) was designed, which concentrates the sampling area in the receptive field corresponding to the target, effectively reducing the interference of complex background information, and enhancing the transmission of important features based on the contribution of different levels to the final target. Secondly, due to the inability of YOLOv7’s detection head to dynamically focus on important features, it is proposed to replace it with a Dynamic Head. This detection head is proficient in capturing feature information across various scales and positions with great effectiveness, and through self-attention, the model can better understand contextual relationships, significantly improving the capacity of the target detection head to accurately represent objects. Ultimately, because the CIoU loss function fails to capture the correlation between the dimensions (width and height) of the predicted box and the desired box, it was replaced with the Focal EIoU loss function to address the previous defects, and Focal EIoU can balance the gradient effects brought by high-quality and low-quality images. After extensive experimental verification on the publicly available VisDrone2019 dataset, our model has achieved significant performance improvements, with $AP_{50}$ increasing by 2.2%, $AP_{\mathbf{S}}, \mathbf{AP} _{\mathbf{M}}$, and $AP_{\mathbf{L}}$increasing by 1.3%, 2.3%, and 3.0%, respectively.

Object detection in remote sensing imagery is a challenging task in the field of computer vision and has high research value. To improve the classification accuracy and positioning accuracy of object detection, we propose a new multi-scale oriented object detector suitable for small objects. Firstly, the feature fusion network based on information balance (IBFF) is proposed to reduce the reuse of different layers’ features from the backbone network and reduce the interference of redundant information based on the premise that the output features have sufficient information, and retain enough shallow detail information. Secondly, to efficiently utilize deep and shallow features, enhance important features, and reduce background noise interference, different attention-based context feature fusion modules (DACFF) are designed according to the characteristics of different feature fusion stages. Finally, an improved strategy of oriented bounding box regression is proposed to obtain the oriented bounding box with a simpler and more effective strategy. The proposed method was evaluated on two public remote sensing datasets, DOTA and HRSC2016, and their mAP values are 80.96% and 95.01%, respectively, which verified the effectiveness of the proposed algorithm.

As the spatial resolution of remote sensing images continues to increase, the complexity of the information they carry also grows. Remote sensing images are characterized by large imaging areas, scattered distributions of similar objects, intricate boundary shapes, and a high density of small objects, all of which pose significant challenges for semantic segmentation tasks. To address these challenges, we propose a Remote Sensing Image Segmentation Network that Integrates Global–Local Multi-Scale Information with Deep and Shallow Features (GLDSFNet). To better handle the wide variations in object sizes and complex boundary shapes, we design a Global–Local Multi-Scale Feature Fusion Module (GLMFM) that enhances segmentation performance by fully leveraging multi-scale information and global context. Additionally, to improve the segmentation of small objects, we propose a Shallow–Deep Feature Fusion Module (SDFFM), which effectively integrates deep semantic information with shallow spatial features through mutual guidance, retaining the advantages of both. Extensive ablation and comparative experiments conducted on two public remote sensing datasets, ISPRS Vaihingen and Potsdam, demonstrate that our proposed GLDSFNet outperforms state-of-the-art methods.

Airport runways, as the core part of airports, belong to vital national infrastructure, and the target detection and segmentation of airport runways in remote sensing images using deep learning methods have significant research value. Most of the existing airport target detection methods based on deep learning rely on horizontal bounding boxes for localization, which often contain irrelevant background information. Moreover, when detecting multiple intersecting airport runways in a single remote sensing image, issues such as false positives and false negatives are apt to occur. To address these challenges, this study proposes an end-to-end remote sensing image airport runway detection and segmentation method based on an improved Mask RCNN (CR-Mask RCNN). The proposed method uses a rotated region generation network instead of a non-rotated region generation network, allowing it to generate rotated bounding boxes that fit the shape of the airport runway more closely, thus avoiding the interference of a large amount of invalid background information brought about by horizontal bounding boxes. Furthermore, the method incorporates an attention mechanism into the backbone feature extraction network to allocate attention to different airport runway feature map scales, which enhances the extraction of local feature information, captures detailed information more effectively, and reduces issues of false positives and false negatives when detecting airport runway targets. The results indicate that, when comparing horizontal bounding boxes with rotated bounding boxes for detecting and segmenting airport runways, the latter are more precise for complex backgrounds. Furthermore, incorporating an attention mechanism enhances the accuracy of airport runway recognition, making it highly effective and practical.