Atrous Spatial Pyramid Pooling Research Articles

COVID-SegNet: encoder-decoder-based architecture for COVID-19 lesion segmentation in chest X-ray.

The coronavirus disease 2019, initially named 2019-nCOV (COVID-19) has been declared a global pandemic by the World Health Organization in March 2020. Because of the growing number of COVID patients, the world's health infrastructure has collapsed, and computer-aided diagnosis has become a necessity. Most of the models proposed for the COVID-19 detection in chest X-rays do image-level analysis. These models do not identify the infected region in the images for an accurate and precise diagnosis. The lesion segmentation will help the medical experts to identify the infected region in the lungs. Therefore, in this paper, a UNet-based encoder-decoder architecture is proposed for the COVID-19 lesion segmentation in chest X-rays. To improve performance, the proposed model employs an attention mechanism and a convolution-based atrous spatial pyramid pooling module. The proposed model obtained 0.8325 and 0.7132 values of the dice similarity coefficient and jaccard index, respectively, and outperformed the state-of-the-art UNet model. An ablation study has been performed to highlight the contribution of the attention mechanism and small dilation rates in the atrous spatial pyramid pooling module.

Segmentation of thyroid glands and nodules in ultrasound images using the improved U-Net architecture

BackgroundIdentifying thyroid nodules’ boundaries is crucial for making an accurate clinical assessment. However, manual segmentation is time-consuming. This paper utilized U-Net and its improved methods to automatically segment thyroid nodules and glands.MethodsThe 5822 ultrasound images used in the experiment came from two centers, 4658 images were used as the training dataset, and 1164 images were used as the independent mixed test dataset finally. Based on U-Net, deformable-pyramid split-attention residual U-Net (DSRU-Net) by introducing ResNeSt block, atrous spatial pyramid pooling, and deformable convolution v3 was proposed. This method combined context information and extracts features of interest better, and had advantages in segmenting nodules and glands of different shapes and sizes.ResultsDSRU-Net obtained 85.8% mean Intersection over Union, 92.5% mean dice coefficient and 94.1% nodule dice coefficient, which were increased by 1.8%, 1.3% and 1.9% compared with U-Net.ConclusionsOur method is more capable of identifying and segmenting glands and nodules than the original method, as shown by the results of correlational studies.

Multi-Scale Ship Detection Algorithm Based on YOLOv7 for Complex Scene SAR Images

Recently, deep learning techniques have been extensively used to detect ships in synthetic aperture radar (SAR) images. The majority of modern algorithms can achieve successful ship detection outcomes when working with multiple-scale ships on a large sea surface. However, there are still issues, such as missed detection and incorrect identification when performing multi-scale ship object detection operations in SAR images of complex scenes. To solve these problems, this paper proposes a complex scenes multi-scale ship detection model, according to YOLOv7, called CSD-YOLO. First, this paper suggests an SAS-FPN module that combines atrous spatial pyramid pooling and shuffle attention, allowing the model to focus on important information and ignore irrelevant information, reduce the feature loss of small ships, and simultaneously fuse the feature maps of ship targets on various SAR image scales, thereby improving detection accuracy and the model’s capacity to detect objects at several scales. The model’s optimization is then improved with the aid of the SIoU loss function. Finally, thorough tests on the HRSID and SSDD datasets are presented to support our methodology. CSD-YOLO achieves better detection performance than the baseline YOLOv7, with a 98.01% detection accuracy, a 96.18% recall, and a mean average precision (mAP) of 98.60% on SSDD. In addition, in comparative experiments with other deep learning-based methods, in terms of overall performance, CSD-YOLO still performs better.

CADNet: an advanced architecture for automatic detection of coronary artery calcification and shadow border in intravascular ultrasound (IVUS) images.

Intravascular Ultrasound (IVUS) is a medical imaging modality widely used for the detection and treatment of coronary heart disease. The detection of vascular structures is extremely important for accurate treatment procedures. Manual detection of lumen and calcification is very time-consuming and requires technical experience. Ultrasound imaging suffers from the generation of artifacts which obstructs the clear delineation among structures. Considering, the need, to provide special attention to crucial areas, convolutional block attention modules (CBAM) is integrated into an encoder-decoder-based U-Net architecture along with Atrous Spatial Pyramid Pooling (ASPP) to detect vessel components: lumen, calcification and shadow borders. The attention modules prove effective in dealing with areas of special attention by assigning additional weights to crucial channels and preserving spatial features. The IVUS data of 12 patients undergoing the treatment is taken for this study. The novelty of the model design is such that it is able to detect the lumen area in the presence/absence of calcification and bifurcation artifacts too. Also, the model efficiently detects the calcification area even in case of severely complex lesions with shadows behind them. The main contribution of the work is that IVUS images of varying degrees of calcification till 360° are also considered in this work, which is usually neglected in previous studies. The experimental results of 1097 IVUS images of 12 patients resulted in meanIoU (0.7894 ± 0.011), Dice Coefficient (0.8763 ± 0.070), precision (0.8768 ± 0.069) and recall (0.8774 ± 0.071) of the proposed model CADNet which show the model's effectiveness relative to other state-of-the art methods.

Wild Mushroom Classification Based on Improved MobileViT Deep Learning

Wild mushrooms are not only tasty but also rich in nutritional value, but it is difficult for non-specialists to distinguish poisonous wild mushrooms accurately. Given the frequent occurrence of wild mushroom poisoning, we propose a new multidimensional feature fusion attention network (M-ViT) combining convolutional networks (ConvNets) and attention networks to compensate for the deficiency of pure ConvNets and pure attention networks. First, we introduced an attention mechanism Squeeze and Excitation (SE) module in the MobilenetV2 (MV2) structure of the network to enhance the representation of picture channels. Then, we designed a Multidimension Attention module (MDA) to guide the network to thoroughly learn and utilize local and global features through short connections. Moreover, using the Atrous Spatial Pyramid Pooling (ASPP) module to obtain longer distance relations, we fused the model features from different layers, and used the obtained joint features for wild mushroom classification. We validated the model on two datasets, mushroom and MO106, and the results showed that M-ViT performed the best on the two test datasets, with accurate dimensions of 96.21% and 91.83%, respectively. We compared the performance of our method with that of more advanced ConvNets and attention networks (Transformer), and our method achieved good results.

RL-DeepLabv3+: A lightweight rice lodging semantic segmentation model for unmanned rice harvester

Lodging is a common natural hazard that occurs in rice during harvest. Currently, the unmanned rice harvester lacks a warning for lodging detection, which often results in harvesting omissions and machine blockages, degrading operational efficiency. Therefore, we proposed a lightweight semantic segmentation model RL-DeepLabv3 + for rice lodging detection of the unmanned rice harvester. We designed a backbone network Rice Lodging-Channel-wise Feature Pyramid (RL-CFP), replaced the Atrous Spatial Pyramid Pooling (ASPP) module with the Channel Attention-based Deep Separable Dilated Convolutional Pyramid (CD-ASP) module, and added the Channel Attention Module (CAM) to the decoding and encoding stages. We trained and evaluated the model using a homemade rice lodging dataset. The model size is 3.39 MB, and the number of parameters is 7.80 × 105. The mean intersection over union (mIoU) and mean pixel accuracy (mPA) of the model were 90.52 % and 94.73 %, which were higher than those of other comparable models. In application tests, when the input image resolutions were 640 × 480 and 1280 × 720, the model detection speeds were 50 frames/s and 21 frames/s in the vision system of the unmanned rice harvester. The proposed model has the advantages of high lodging detection accuracy, small model space occupation, and uncomplicated application deployment. It can provide a warning to the unmanned rice harvester for lodging detection and improve its intelligence level.

Three-dimensional nanoscale reduced-angle ptycho-tomographic imaging with deep learning (RAPID)

X-ray ptychographic tomography is a nondestructive method for three dimensional (3D) imaging with nanometer-sized resolvable features. The size of the volume that can be imaged is almost arbitrary, limited only by the penetration depth and the available scanning time. Here we present a method that rapidly accelerates the imaging operation over a given volume through acquiring a limited set of data via large angular reduction and compensating for the resulting ill-posedness through deeply learned priors. The proposed 3D reconstruction method “RAPID” relies initially on a subset of the object measured with the nominal number of required illumination angles and treats the reconstructions from the conventional two-step approach as ground truth. It is then trained to reproduce equal fidelity from much fewer angles. After training, it performs with similar fidelity on the hitherto unexamined portions of the object, previously not shown during training, with a limited set of acquisitions. In our experimental demonstration, the nominal number of angles was 349 and the reduced number of angles was 21, resulting in a times 140 aggregate speedup over a volume of 4.48times 93.18times 3.92, upmu text {m}^3 and with (14,text {nm})^3 feature size, i.e. sim 10^8 voxels. RAPID’s key distinguishing feature over earlier attempts is the incorporation of atrous spatial pyramid pooling modules into the deep neural network framework in an anisotropic way. We found that adjusting the atrous rate improves reconstruction fidelity because it expands the convolutional kernels’ range to match the physics of multi-slice ptychography without significantly increasing the number of parameters.

Lightweight Semantic segmentation method based on GhostNet and Atrous Spatial Pyramid Pooling

Semantic segmentation is frequently utilized in computer vision projects like remote sensing picture segmentation, unmanned driving, and medical image segmentation. A lightweight semantic segmentation model is suggested by taking into account three components of the network - parameters, calculation, and performance - to address the issue of deploying embedded platforms with constrained processing power and hardware storage. Dual Attention is used with Atrous Spatial Pyramid Pooling (ASPP) to obtain precise relevant data utilizing the lightweight network GhostNet as the foundation, and to lighten ASPP’s computational burden, depthwise separable convolution is used. To obtain a distinct segmentation boundary, a multi-scale splicing technique is then used. With network parameters of 2.78106, a floating-point calculation of 1.931GFLOPs, and a MIoU of 72.13% in the experiments just on PASCAL VOC 2012, the model achieves an excellent compromise between computational efficiency and segmentation precision.

Li-SegPNet: Encoder-decoder Mode Lightweight Segmentation Network for Colorectal Polyps Analysis.

One of the fundamental and crucial tasks for the automated diagnosis of colorectal cancer is the segmentation of the acute gastrointestinal lesions, most commonly colorectal polyps. Therefore, in this work, we present a novel lightweight encoder-decoder mode of architecture with the attention mechanism to address this challenging task. The proposed Li-SegPNet architecture harnesses cross-dimensional interaction in feature maps with novel encoder block with modified triplet attention. We have used atrous spatial pyramid pooling to handle the problem of segmenting objects at multiple scales. We also address the semantic gap between the encoder and decoder through a modified skip connection using attention gating. We applied our model to colonoscopy still images and trained and validated it on two publicly available datasets, Kvasir-SEG and CVC-ClinicDB. We achieve mean Intersection-Over-Union (mIoU) and dice scores of 0.88, 0.9058 and 0.8969, 0.9372 on Kvasir-SEG and CVC-ClinicDB, respectively. We analyze the generalizability of Li-SegPNet by testing it on two independent previously unseen datasets, Hyper-Kvasir and EndoTect 2020, and establish the model efficiency in cross-dataset evaluation. We employ multi-scale testing to examine the model performance on different sizes of polyps. Li-SegPNet performs best on medium-sized polyps with a mIoU and dice score of 0.9086 and 0.9137, respectively on the Kvasir-SEG dataset and 0.9425, 0.9434 of mIoU and dice score, respectively on CVC-ClinicDB. The experimental results convey that we establish a new benchmark on these four datasets for the segmentation of polyps. The proposed model can be used as a new benchmark model for polyps segmentation. Lesser parameters in comparison to other models give the edge in the applicability of the proposed Li-SegPNet model in real-time clinical analysis.

RSCDNet: A Robust Deep Learning Architecture for Change Detection From Bi-Temporal High Resolution Remote Sensing Images

Accurate change detection from high-resolution satellite and aerial images is of great significance in remote sensing for precise comprehension of Land cover (LC) variations. The current methods compromise with the spatial context; hence, they fail to detect and delineate small change areas and are unable to capture the difference between features of the bi-temporal images. This paper proposes Remote Sensing Change Detection Network (RSCDNet) - a robust end-to-end deep learning architecture for pixel-wise change detection from bi-temporal high-resolution remote-sensing (HRRS) images. The proposed RSCDNet model is based on an encoder-decoder framework integrated with the Modified Self-Attention (MSA) andthe Gated Linear Atrous Spatial Pyramid Pooling (GL-ASPP) blocks; both efficient mechanisms to regulate the field-of-view while finding the most suitable trade-off between accurate localization and context assimilation. The paper documents the design and development of the proposed RSCDNet model and compares its qualitative and quantitative results with state-of-the-art HRRS change detection architectures. The above mentioned novelties in the proposed architecture resulted in an F1-score of 98%, 98%, 88%, and 75% on the four publicly available HRRS datasets namely, Staza-Tisadob, Onera, CD-LEVIR, and WHU. In addition to the improvement in the performance metrics, the strategic connections in the proposed GL-ASPP and MSA units significantly reduce the prediction time per image (PTPI) and provide robustness against perturbations. Experimental results yield that the proposed RSCDNet model outperforms the most recent change detection benchmark models on all four HRRS datasets.

Adaptive Context Exploration Network for Polyp Segmentation in Colonoscopy Images

Recently, automatic and accurate polyp segmentation has become an emerging yet challenging issue. Although convolutional neural networks (CNNs) exhibit a promising future modality to address this issue, most CNN-based methods highly require extensive labeled data. Unfortunately, there is a lack of large-scale public colorectal polyp segmentation datasets in the clinical community and academia. In this study, we construct a new benchmark dataset, which includes 2163 colonoscopy images and their pixel-wise annotations. Moreover, for intelligent polyp segmentation, we propose a novel adaptive context exploration network (ACENet). Our ACENet follows an encoder-decoder architecture and consists of two key modules, i.e., an attentional atrous spatial pyramid pooling (AASPP) module and an adaptive context extraction (ACE) module. The AASPP fuses semantic features from the encoder, and generates the global guidance information for the following decoder. The ACE captures multi-scale features and aggregates them by a branch-wise attention mechanism. Benefiting from these two modules, our ACENet is capable of adaptively exploring the context features to locate and detect the polyp regions effectively. Extensive experiments on the collected dataset and four publicly available datasets show that the proposed ACENet achieves superior performance on five evaluation metrics over three mainstream categories of the state-of-the-art methods.

High-turbidity underwater active single-pixel imaging based on generative adversarial networks with double Attention U-Net under low sampling rate

Single-pixel imaging (SPI) is widely used in underwater optical imaging because it has the feature of turbulence-free, little backscattering, and simple system structure. However, the image reconstruction quality of SPI is severely degraded as the increasing intensity of scattering medium in underwater environments. In this paper, an underwater active SPI system based on generative adversarial network with double Attention U-Net is proposed to reconstruct the underwater target image with high turbidity under low sampling rates. The generator of the proposed network use two modified Attention U-Net with double skip connections between different layers to improve the fidelity of reconstructed images. The atrous spatial pyramid pooling, and spatial squeeze and channel excitation modules are also integrated to obtain multi-scale information and remove redundant information. What is more, the least squares loss, content loss and mean structural similarity loss are also incorporated into the total loss function to stabilize the training process and avoid the gradient disappearance. Numerical simulations and physical experiments have shown that the proposed method can reconstruct underwater target images at a low sampling rate of 3.52% under the turbidity of 128NTU. The proposed method shows relatively strong reconstruction ability and generalization, and the PSNR and SSIM can be respectively improved by 0.75 times and 3 times compared with compressed sensing SPI. Our work provides a new insight into underwater SPI with high turbidity and can greatly improve the capability of underwater SPI.

Semantic segmentation of remote sensing images based on dilated convolution and spatial-channel attention mechanism

The rich context information and multiscale ground information in remote sensing images are crucial to improving the semantic segmentation accuracy. Therefore, we propose a remote sensing image semantic segmentation method that integrates multilevel spatial channel attention and multi-scale dilated convolution, effectively addressing the issue of poor segmentation performance of small target objects in remote sensing images. This method builds a multilevel characteristic fusion structure, combining deep-level semantic characteristics with the details of the shallow levels to generate multiscale feature diagrams. Then, we introduce the dilated convolution of the series combination in each layer of the atrous spatial pyramid pooling structure to reduce the loss of small target information. Finally, using convolutional conditional random field to describe the context information on the space and edges to improve the model’s ability to extract details. We prove the effectiveness of the model on the three public datasets. From the quantitative point of view, we mainly evaluate the four indicators of the model’s F1 score, overall accuracy (OA), Intersection over Union (IoU), and Mean Intersection over Union (MIoU). On GID dataset, F1 score, OA, and MIoU reach 87.27, 87.80, and 77.70, respectively, superior to most mainstream semantic segmentation networks.

Contour-enhanced densely connected Siamese network for change detection

Obtaining change information in different periods from a pair of registered satellite remote sensing images is of great significance to urban planning, so change detection (CD) technology has attracted extensive attention in recent years. In recent years, convolutional neural networks have set off a boom in many artificial intelligence research fields because of their excellent feature extraction performance. However, the common convolution operation mainly focuses on the abstraction of the semantic information of the features, which often leads to the details of the features being ignored and thus affects the final accuracy. For example, the contour details of changing objects and the structural information of small objects are often lost. We propose a Siamese network that enhances contour and structural details to achieve higher-accuracy CD tasks for bitemporal remote sensing images. In this network, we propose an efficient contour-enhanced convolutional block that is based on the reparameterization technique. The contour-enhanced convolutional block strengthens the extraction of structural and contour features by integrating different branches. In addition, inspired by NestedUNet and to better preserve the original location information of features, we use a dense connection as the feature extractor to obtain refined features of bitemporal images. After that, we use a difference module to calculate the change characteristics of the dual-time image, and we use atrous spatial pyramid pooling and enhanced spatial attention to further refine the obtained change characteristics. We conduct extensive experiments on three different datasets to verify the effectiveness of our model. Experimental results show that our method outperforms state-of-the-art methods in both overall accuracy and visualization details.

Radar Emitter Identification with Multi-View Adaptive Fusion Network (MAFN)

Radar emitter identification (REI) aims to extract the fingerprint of an emitter and determine the individual to which it belongs. Although many methods have used deep neural networks (DNNs) for an end-to-end REI, most of them only focus on a single view of signals, such as spectrogram, bi-spectrum, signal waveforms, and so on. When the electromagnetic environment varies, the performance of DNN will be significantly degraded. In this paper, a multi-view adaptive fusion network (MAFN) is proposed by simultaneously exploring the signal waveform and ambiguity function (AF). First, the original waveform and ambiguity function of the radar signals are used separately for feature extraction. Then, a multi-scale feature-level fusion module is constructed for the fusion of multi-view features from waveforms and AF, via the Atrous Spatial Pyramid Pooling (ASPP) structure. Next, the class probability is modeled as Dirichlet distribution to perform adaptive decision-level fusion via evidence theory. Extensive experiments are conducted on two datasets, and the results show that the proposed MAFN can achieve accurate classification of radar emitters and is more robust than its counterparts.

RAD-UNet: Research on an improved lung nodule semantic segmentation algorithm based on deep learning.

Due to the small proportion of target pixels in computed tomography (CT) images and the high similarity with the environment, convolutional neural network-based semantic segmentation models are difficult to develop by using deep learning. Extracting feature information often leads to under- or oversegmentation of lesions in CT images. In this paper, an improved convolutional neural network segmentation model known as RAD-UNet, which is based on the U-Net encoder-decoder architecture, is proposed and applied to lung nodular segmentation in CT images. The proposed RAD-UNet segmentation model includes several improved components: the U-Net encoder is replaced by a ResNet residual network module; an atrous spatial pyramid pooling module is added after the U-Net encoder; and the U-Net decoder is improved by introducing a cross-fusion feature module with channel and spatial attention. The segmentation model was applied to the LIDC dataset and a CT dataset collected by the Affiliated Hospital of Anhui Medical University. The experimental results show that compared with the existing SegNet [14] and U-Net [15] methods, the proposed model demonstrates better lung lesion segmentation performance. On the above two datasets, the mIoU reached 87.76% and 88.13%, and the F1-score reached 93.56% and 93.72%, respectively. Conclusion: The experimental results show that the improved RAD-UNet segmentation method achieves more accurate pixel-level segmentation in CT images of lung tumours and identifies lung nodules better than the SegNet [14] and U-Net [15] models. The problems of under- and oversegmentation that occur during segmentation are solved, effectively improving the image segmentation performance.

The Circular U-Net with Attention Gate for Image Splicing Forgery Detection

With the advent and rapid development of image tampering technology, it has become harmful to many aspects of our society. Thus, image tampering detection has been increasingly important. Although current forgery detection methods have achieved some success, the scale of the tampered areas in each forgery image are different, and previous methods do not take this into account. In this paper, we believe that the inability of the network to accommodate tampered regions of various sizes is the main reason for the low precision. To address the mentioned problem, we propose a neural network architecture called CAU-Net, which adds residual propagation and feedback, attention gate and Atrous Spatial Pyramid Pooling with CBAM to the U-Net. The Atrous Spatial Pyramid Pooling with CBAM can capture information from multiple scales and adapt to differently sized target areas. In addition, CAU-Net can solve the vanishing gradient issue and suppress the weight of untampered regions, and CAU-Net is an end-to-end network without redundant image processing; thus, it is fast to detect suspicious images. In the end, we optimize the proposed network structure by ablation study, and the experimental results and visualization results demonstrate that our network has a better performance on CASIA and NIST16 compared with state of the art methods.

Leveraging Deep Convolutional Neural Network for Point Symbol Recognition in Scanned Topographic Maps

Point symbols on a scanned topographic map (STM) provide crucial geographic information. However, point symbol recognition entails high complexity and uncertainty owing to the stickiness of map elements and singularity of symbol structures. Therefore, extracting point symbols from STMs is challenging. Currently, point symbol recognition is performed primarily through pattern recognition methods that have low accuracy and efficiency. To address this problem, we investigated the potential of a deep learning-based method for point symbol recognition and proposed a deep convolutional neural network (DCNN)-based model for this task. We created point symbol datasets from different sources for training and prediction models. Within this framework, atrous spatial pyramid pooling (ASPP) was adopted to handle the recognition difficulty owing to the differences between point symbols and natural objects. To increase the positioning accuracy, the k-means++ clustering method was used to generate anchor boxes that were more suitable for our point symbol datasets. Additionally, to improve the generalization ability of the model, we designed two data augmentation methods to adapt to symbol recognition. Experiments demonstrated that the deep learning method considerably improved the recognition accuracy and efficiency compared with classical algorithms. The introduction of ASPP in the object detection algorithm resulted in higher mean average precision and intersection over union values, indicating a higher recognition accuracy. It is also demonstrated that data augmentation methods can alleviate the cross-domain problem and improve the rotation robustness. This study contributes to the development of algorithms and the evaluation of geographic elements extracted from STMs.

Sci-Net: scale-invariant model for buildings segmentation from aerial imagery

Buildings’ segmentation is a fundamental task in the field of earth observation and aerial imagery analysis. Most existing deep learning-based methods in the literature can be applied to a fixed or narrow-range spatial resolution imagery. In practical scenarios, users deal with a broad spectrum of image resolutions. Thus, a given aerial image often needs to be re-sampled to match the spatial resolution of the dataset used to train the deep learning model, which results in a degradation in segmentation performance. To overcome this challenge, we propose, in this manuscript, scale-invariant neural network (Sci-Net) architecture that segments buildings from wide-range spatial resolution aerial images. Specifically, our approach leverages UNet hierarchical representation and dense atrous spatial pyramid pooling to extract fine-grained multi-scale representations. Sci-Net significantly outperforms state-of-the-art models on the open cities AI and the multi-scale building datasets with a steady improvement margin across different spatial resolutions.

Feature enhancement and supervised contrastive learning for image splicing forgery detection

Image forgery detection remains a challenging task due to the variation in the scale of the tampered areas and the forensic clues that are obscured by various post-processing operations. Most existing deep learning methods rely only on single-scale high-level features and ignore the correlations of the pixels in intra-image and inter-image. Due to this limitation, previous methods are considered unsuitable for multi-scale splicing forgery detection. To fill this gap, we propose a novel model for improving multi-scale splicing forged regions localization by utilizing multi-level multi-scale feature enhancement and pixel-level supervised contrastive learning. First, we introduce a multi-level multi-scale feature enhancement module (MFEM) to integrate the multi-level information and capture the multi-scale global contextual representation by embedding an improved atrous spatial pyramid pooling (ASPP) mechanism into the non-local module. It strengthens the capability of the model to sense multi-scale tampered regions. Second, the pixel-level supervised contrastive learning mechanism is designed to separately cluster the pixel representations of tampered and real regions within and across images. It improves intra-class compactness and inter-class separability of the pixel embedding space significantly and enhances feature expression capabilities. Third, we design a multi-loss progressive learning (MPL) strategy to integrate the complementary advantages of multi-loss functions to optimize the scale and position parameters of the tampered regions during the training process. Extensive experiments have shown that the proposed model outperforms state-of-the-art methods. It can effectively detect and segment multi-scale tampered regions, even for noisy and JPEG-compressed images.