Deep Learning-based Detection Research Articles

Enhanced Disease Detection for Apple Leaves with Rotating Feature Extraction

Leaf diseases such as Mosaic disease and Black Rot are among the most common diseases affecting apple leaves, significantly reducing apple yield and quality. Detecting leaf diseases is crucial for the prevention and control of these conditions. In this paper, we propose incorporating rotated bounding boxes into deep learning-based detection, introducing the ProbIoU loss function to better quantify the difference between model predictions and real results in practice. Specifically, we integrated the Plant Village dataset with an on-site dataset of apple leaves from an orchard in Weifang City, Shandong Province, China. Additionally, data augmentation techniques were employed to expand the dataset and address the class imbalance issue. We utilized the EfficientNetV2 architecture with inverted residual structures (FusedMBConv and S-MBConv modules) in the backbone network to build sparse features using a top–down approach, minimizing information loss. The inclusion of the SimAM attention mechanism effectively captures both channel and spatial attention, expanding the receptive field and enhancing feature extraction. Furthermore, we introduced depth-wise separable convolution and the CAFM in the neck network to improve feature fusion capabilities. Finally, experimental results demonstrate that our model outperforms other detection models, achieving 93.3% mAP@0.5, 88.7% Precision, and 89.6% Recall. This approach provides a highly effective solution for the early detection of apple leaf diseases, with the potential to significantly improve disease management in apple orchards.

Clinical Utility of Deep Learning Assistance for Detecting Various Abnormal Findings in Color Retinal Fundus Images: A Reader Study.

To evaluate the clinical usefulness of a deep learning-based detection device for multiple abnormal findings on retinal fundus photographs for readers with varying expertise. Fourteen ophthalmologists (six residents, eight specialists) assessed 399 fundus images with respect to 12 major ophthalmologic findings, with or without the assistance of a deep learning algorithm, in two separate reading sessions. Sensitivity, specificity, and reading time per image were compared. With algorithmic assistance, readers significantly improved in sensitivity for all 12 findings (P < 0.05) but tended to be less specific (P < 0.05) for hemorrhage, drusen, membrane, and vascular abnormality, more profoundly so in residents. Sensitivity without algorithmic assistance was significantly lower in residents (23.1%∼75.8%) compared to specialists (55.1%∼97.1%) in nine findings, but it improved to similar levels with algorithmic assistance (67.8%∼99.4% in residents, 83.2%∼99.5% in specialists) with only hemorrhage remaining statistically significantly lower. Variances in sensitivity were significantly reduced for all findings. Reading time per image decreased in images with fewer than three findings per image, more profoundly in residents. When simulated based on images acquired from a health screening center, average reading time was estimated to be reduced by 25.9% (from 16.4 seconds to 12.1 seconds per image) for residents, and by 2.0% (from 9.6 seconds to 9.4 seconds) for specialists. Deep learning-based computer-assisted detection devices increase sensitivity, reduce inter-reader variance in sensitivity, and reduce reading time in less complicated images. This study evaluated the influence that algorithmic assistance in detecting abnormal findings on retinal fundus photographs has on clinicians, possibly predicting its influence on clinical application.

Deep Learning-Based Detection of Carotid Plaques Informs Cardiovascular Risk Prediction and Reveals Genetic Drivers of Atherosclerosis.

Atherosclerotic cardiovascular disease, the leading cause of global mortality, is driven by lipid accumulation and plaque formation within arterial walls. Carotid plaques, detectable via ultrasound, are a well-established marker of subclinical atherosclerosis. In this study, we trained a deep learning model to detect plaques in 177,757 carotid ultrasound images from 19,499 UK Biobank (UKB) participants (aged 47-83 years) to assess the prevalence, risk factors, prognostic significance, and genetic architecture of carotid atherosclerosis in a large population-based cohort. The model demonstrated high performance metrics with accuracy, sensitivity, specificity, and positive predictive value of 89.3%, 89.5%, 89.2%, and 82.9%, respectively, identifying carotid plaques in 45% of the population. Plaque presence and count were significantly associated with future cardiovascular events over a median follow-up period of up to 7 years, leading to improved risk reclassification beyond established clinical prediction models. A genome-wide association study (GWAS) meta-analysis of carotid plaques (29,790 cases, 36,847 controls) uncovered two novel genomic loci (p < 5×10 -8 ) with downstream analyses implicating lipoprotein(a) and interleukin-6 signaling, both targets of investigational drugs in advanced clinical development. Observational and Mendelian randomization analyses showed associations between smoking, low-density-lipoprotein (LDL) cholesterol, and high blood pressure and the odds of carotid plaque presence. Our study underscores the potential of carotid plaque assessment for improving cardiovascular risk prediction, provides novel insights into the genetic basis of subclinical atherosclerosis, and offers a valuable resource for advancing atherosclerosis research at the population scale.

Deep learning-based detection of affected body parts in Parkinson’s disease and freezing of gait using time-series imaging

We proposed a deep learning method using a convolutional neural network on time-series (TS) images to detect and differentiate affected body parts in people with Parkinson’s disease (PD) and freezing of gait (FOG) during 360° turning tasks. The 360° turning task was performed by 90 participants (60 people with PD [30 freezers and 30 nonfreezers] and 30 age-matched older adults (controls) at their preferred speed. The position and acceleration underwent preprocessing. The analysis was expanded from temporal to visual data using TS imaging methods. According to the PD vs. controls classification, the right lower third of the lateral shank (RTIB) on the least affected side (LAS) and the right calcaneus (RHEE) on the LAS were the most relevant body segments in the position and acceleration TS images. The RHEE marker exhibited the highest accuracy in the acceleration TS images. The identified markers for the classification of freezers vs. nonfreezers vs. controls were the left lateral humeral epicondyle (LELB) on the more affected side and the left posterior superior iliac spine (LPSI). The LPSI marker in the acceleration TS images displayed the highest accuracy. This approach could be a useful supplementary tool for determining PD severity and FOG.

TAENet: Two-branch Autoencoder Network for Interpretable Deepfake Detection

Deepfake detection attracts increasingly attention due to serious security issues caused by facial manipulation techniques. Recently, deep learning-based detectors have achieved promising performance. However, these detectors suffer severe untrustworthy due to the lack of interpretability. Thus, it is essential to work on the interpretibility of deepfake detectors to improve the reliability and traceability of digital evidence. In this work, we propose a two-branch autoencoder network named TAENet for interpretable deepfake detection. TAENet is composed of Content Feature Disentanglement (CFD), Content Map Generation (CMG), and Classification. CFD extracts latent features of real and forged content with dual encoder and feature discriminator. CMG employs a Pixel-level Content Map Generation Loss (PCMGL) to guide the dual decoder in visualizing the latent representations of real and forged contents as real-map and fake-map. In classification module, the Auxiliary Classifier (AC) serves as map amplifier to improve the accuracy of real-map image extraction. Finally, the learned model decouples the input image into two maps that have the same size as the input, providing visualized evidence for deepfake detection. Extensive experiments demonstrate that TAENet can offer interpretability in deepfake detection without compromising accuracy.

Deep learning-based detection and classification of intracranial tumors on magnetic resonance imaging

AbstractBackgroundWe evaluated the capability of an AI application to independently detect, segment and classify intracranial tumors in MRI.MethodsIn this retrospective single-centre study, 138 patients (65 female and 73 male) with a mean age of 35 ± 26y were included. 97 were diagnosed with an intracranial neoplasm, while 41 exhibited no intracranial pathology. Inclusion criteria were a 1.5 or 3.0T MRI dataset with the following sequences: T2 axial, T1 axial pre- and post-contrast with a slice thickness between 3 and 6 mm and no previous brain surgery. Image analysis was performed by two human readers (R1 = 5 years and R2 = 10 years of experience in brain MRI) and a deep learning (DL)-based AI model. Sensitivity, specificity and accuracy of the AI model and the human readers to detect and correctly classify brain tumors were measured. Histological results served as the gold standard.ResultsThe AI model reached a sensitivity of 93.81% [87.02–97.70] and a specificity of 63.41% [46.94–77.88], while human readers reached 100% [96.27–100.00] and 100% [91.40–100.00], respectively. Human readers provided a significantly higher accuracy rate with R1 0.93 (95% CI: 0.88–0.97) and R2 0.98 (95% CI: 0.94, 0.99) vs. 0.74 (95% CI: 0.66–0.81) for the AI model (P-value &lt;0.001).ConclusionThe underlying DL-based AI algorithm can independently identify and segment intracranial tumors while providing satisfactory results for establishing the correct diagnosis. Despite its current inferiority compared to experienced radiologists, it still experiences ongoing development and it is a step towards developing an artificial intelligence-augmented radiology workflow.

Deep Learning-Based Detection of Reticular Pseudodrusen in Age-Related Macular Degeneration on Optical Coherence Tomography.

Reticular pseudodrusen (RPD) signify a critical phenotype driving vision loss in age-related macular degeneration (AMD). Their detection is paramount in the clinical management of those with AMD, yet they remain challenging to reliably identify. We thus developed a deep learning (DL) model to segment RPD from 9,800 optical coherence tomography B-scans, and this model produced RPD segmentations that had higher agreement with four retinal specialists (Dice similarity coefficient [DSC]=0·76 [95% confidence interval [CI] 0·71-0·81]) than the agreement amongst the specialists (DSC=0·68, 95% CI=0·63-0·73; p <0·001). In five external test datasets consisting of 1,017 eyes from 812 individuals, the DL model detected RPD with a similar level of performance as two retinal specialists (area-under-the-curve of 0·94 [95% CI=0·92-0·97], 0·95 [95% CI=0·92-0·97] and 0·96 [95% CI=0·94-0·98] respectively; p ≥0·32). This DL model enables the automatic detection and quantification of RPD with expert-level performance, which we have made publicly available.

Deep learning-based detection of qanat underground water distribution systems using HEXAGON spy satellite imagery

Qanats are a remarkable type of ancient hydraulic structure for sustainable water distribution in arid environments that use subterranean channels to transport water from highland or mountainous areas. The presence of the qanat system is marked by a line of regularly spaced shafts visible from the surface, which can be used to detect qanats using satellite imagery. Typically, qanats have been documented by field mapping or manual digitisation within a Geographic Information System (GIS) environment. This process is time-consuming due to the numerous shafts within each qanat line. However, several automated methods for detecting qanat structures have been explored, using techniques such as morphological filters, custom convolutional neural networks (CNN) and, more recently, YOLOv5 and Mask R-CNN. These approaches used high-resolution RGB images and CORONA images. However, the use of black and white CORONA in CNNs has been limited in its applicability due to a high rate of false positives.This paper explores the potential of YOLOv9 in processing the black and white HEXAGON (KH-9) high-resolution spy satellite system launched in 1971. Two areas in Afghanistan (Maiwand) and Iran (Gorgan Plain) were selected to train the system images extracted from HEXAGON imagery and artificial synthetic data. The training dataset was augmented using the Albumentation library, which increased the number of tiles used. The model was tested using two types of HEXAGON imagery for selected areas in Afghanistan (Maiwand), Iran (Gorgan Plain) and Morocco (Rissani), and CORONA imagery in Iran (Gorgan Plain).Our study provided a model capable of predicting the location of qanat shafts with a precision of over 0.881 and a recall of 0.627 for most of the case studies tested. This is the first case study aimed at detecting qanats in different landscapes using different types of satellite imagery. Using real, augmented, and artificial data allowed us to generalise the representation of qanats into lineal groups of circular features. Thanks to applying labelling for individual qanats and their pairs as separate classes, our approach eliminated most of the isolated and clustered false positives.

Charge Manipulation Attacks Against Smart Electric Vehicle Charging Stations and Deep Learning-Based Detection Mechanisms

Charge Manipulation Attacks Against Smart Electric Vehicle Charging Stations and Deep Learning-Based Detection Mechanisms

Enhancing Periodontal Treatment Through the Integration of Deep Learning-Based Detection with Bayesian Network Models.

This study incorporated deep learning for periodontal disease detection into a Bayesian network (BN) clinical decision support model for comprehensive periodontal care. BN structure and probabilities were based on clinical data and Faster R-CNN-detected radiographic images. Receiver operating characteristic curve analysis confirmed the model's high accuracy in treatment plan recommendations.

A cosine similarity-based labeling technique for vulnerability type detection using source codes

Vulnerability detection is of great importance in providing reliability to software systems. Although existing methods achieve remarkable success in vulnerability detection, they have several disadvantages as follows: (1) The irrelevant information is removed from source codes, which have a high noise ratio, thereby utilizing deep learning methods and devising experiments featuring high accuracy. However, deep learning-based detection methods necessitate large-scale datasets. This results in computational hardship with respect to vulnerability detection in small-scale software systems. (2) The majority of the studies perform feature selection by processing vulnerability commits. Despite tremendous endeavors, there are few works detecting vulnerability with source codes. To solve these two problems, in this study, a novel labeling and vulnerability detection algorithm is proposed. The algorithm first exploits source codes with the help of a keyword vulnerability matrix. After that, an ultimate encoded matrix is generated by word2vec, thereby combining the labeling vector with the source code matrix to reveal a trainable dataset for a generalized linear model (GLM). Different from preceding studies, our method performs vulnerability detection without requiring vulnerability commits but using source codes. In addition to this, similar studies generally aim to bring sophisticated solutions for just one type of programming language. Conversely, our study develops vulnerability keywords for three programming languages including C#, Java, and C++, and creates the related labeling vectors by regarding the keyword matrix. The proposed method outperformed the baseline approaches for most of the experimental datasets with over 90% of the area under the curve (AUC). Further, there is a 7.7% margin between our method and the alternatives on average for Recall, Precision, and F1-score with respect to five types of vulnerabilities.

Distracted Driving Behavior Detection Algorithm Based on Lightweight StarDL-YOLO

Distracted driving is one of the major factors leading drivers to ignore potential road hazards. In response to the challenges of high computational complexity, limited generalization capacity, and suboptimal detection accuracy in existing deep learning-based detection algorithms, this paper introduces a novel approach called StarDL-YOLO (StarNet-detectlscd-yolo), which leverages an enhanced version of YOLOv8n. Initially, the StarNet integrated into the backbone of YOLOv8n significantly improves the feature extraction capability of the model with remarkable reduction in computational complexity. Subsequently, the Star Block is incorporated into the neck network, forming a C2f-Star module that offers lower computational cost. Additionally, shared convolution is introduced in the detection head to further reduce computational burden and parameter size. Finally, the Wise-Focaler-MPDIoU loss function is proposed to strengthen detection accuracy. The experimental results demonstrate that StarDL-YOLO significantly improves the efficiency of the distracted driving behavior detection, achieving an accuracy of 99.6% on the StateFarm dataset. Moreover, the parameter count of the model is minimized by 56.4%, and its computational load is decreased by 45.1%. Additionally, generalization experiments are performed on the 100-Driver dataset, revealing that the proposed scheme enhances generalization effectiveness compared to YOLOv8n. Therefore, this algorithm significantly reduces computational load while maintaining high reliability and generalization capability.

Deep learning-based detection and semi-quantitative model for spread through air spaces (STAS) in lung adenocarcinoma

Tumor spread through air spaces (STAS) is a distinctive metastatic pattern affecting prognosis in lung adenocarcinoma (LUAD) patients. Several challenges are associated with STAS detection, including misdetection, low interobserver agreement, and lack of quantitative analysis. In this research, a total of 489 digital whole slide images (WSIs) were collected. The deep learning-based STAS detection model, named STASNet, was constructed to calculate semi-quantitative parameters associated with STAS density and distance. STASNet demonstrated an accuracy of 0.93 for STAS detection at the tiles level and had an AUC of 0.72–0.78 for determining the STAS status at the WSI level. Among the semi-quantitative parameters, T10S, combined with the spatial location information, significantly stratified stage I LUAD patients on disease-free survival. Additionally, STASNet was deployed into a real-time pathological diagnostic environment, which boosted the STAS detection rate and led to the identification of three easily misidentified types of occult STAS.

Deep Learning-Based Detection and Robust Tracking of an Unmanned Surface Ship Using a Monocular Camera

Deep Learning-Based Detection and Robust Tracking of an Unmanned Surface Ship Using a Monocular Camera

Deep learning-based detection of primary bone tumors around the knee joint on radiographs: a multicenter study.

Most primary bone tumors are often found in the bone around the knee joint. However, the detection of primary bone tumors on radiographs can be challenging for the inexperienced or junior radiologist. This study aimed to develop a deep learning (DL) model for the detection of primary bone tumors around the knee joint on radiographs. From four tertiary referral centers, we recruited 687 patients diagnosed with bone tumors (including osteosarcoma, chondrosarcoma, giant cell tumor of bone, bone cyst, enchondroma, fibrous dysplasia, etc.; 417 males, 270 females; mean age 22.8±13.2 years) by postoperative pathology or clinical imaging/follow-up, and 1,988 participants with normal bone radiographs (1,152 males, 836 females; mean age 27.9±12.2 years). The dataset was split into a training set for model development, an internal independent and an external test set for model validation. The trained model located bone tumor lesions and then detected tumor patients. Receiver operating characteristic curves and Cohen's kappa coefficient were used for evaluating detection performance. We compared the model's detection performance with that of two junior radiologists in the internal test set using permutation tests. The DL model correctly localized 94.5% and 92.9% bone tumors on radiographs in the internal and external test set, respectively. An accuracy of 0.964/0.920, and an area under the receiver operating characteristic curve (AUC) of 0.981/0.990 in DL detection of bone tumor patients were for the internal and external test set, respectively. Cohen's kappa coefficient of the model in the internal test set was significantly higher than that of the two junior radiologists with 4 and 3 years of experience in musculoskeletal radiology (Model vs. Reader A, 0.927 vs. 0.777, P<0.001; Model vs. Reader B, 0.927 vs. 0.841, P=0.033). The DL model achieved good performance in detecting primary bone tumors around the knee joint. This model had better performance than those of junior radiologists, indicating the potential for the detection of bone tumors on radiographs.

Deep Learning-Based Detection of Pituitary, Glioma, and Meningioma Tumors from Brain MRIs

Deep Learning-Based Detection of Pituitary, Glioma, and Meningioma Tumors from Brain MRIs

Assessing the Effectiveness of Network Security Tools in Mitigating the Impact of Deepfakes AI on Public Trust in Media

The rising threat of deepfake technology challenges public trust in media, necessitating robust countermeasures. This study proposes the Anti-DFK framework, a comprehensive strategy to mitigate the spread of deepfakes on major social platforms such as Instagram, Facebook, YouTube, and Twitter. The framework integrates deep learning-based detection engines, digital watermarking, and advanced network access controls, including URL filtering, domain reputation filtering, content-type filtering, and Geo-IP blocking. Analyzing historical deepfake data, user engagement metrics, and public sentiment from Kaggle Datasets, the study employed deep learning models—CNNs, LSTMs, and Transformer-based—to evaluate detection capabilities, achieving the highest controlled environment accuracy of 0.97. Digital watermarking techniques were tested for robustness against various attacks, with the DCT method displaying significant resilience. Network access controls were assessed for their effectiveness in curtailing the spread of deepfakes, with content filtering proving the most effective by reducing dissemination by nearly 80%. Findings indicate a critical negative impact of deepfakes on public trust, underscoring the need for the integrated approach offered by the Anti-DFK framework. The study concludes that implementing these sophisticated detection tools, combined with robust digital watermarking and stringent network controls, can significantly enhance the integrity of media content and restore public confidence.

SPA: Annotating Small Object with a Single Point in Remote Sensing Images

Detecting oriented small objects is a critical task in remote sensing, but the development of high-performance deep learning-based detectors is hindered by the need for large-scale and well-annotated datasets. The high cost of creating these datasets, due to the dense and numerous distribution of small objects, significantly limits the application and development of such detectors. To address this problem, we propose a single-point-based annotation approach (SPA) based on the graph cut method. In this framework, user annotations act as the origin of positive sample points, and a similarity matrix, computed from feature maps extracted by deep learning networks, facilitates an intuitive and efficient annotation process for building graph elements. Utilizing the Maximum Flow algorithm, SPA derives positive sample regions from these points and generates oriented bounding boxes (OBBOXs). Experimental results demonstrate the effectiveness of SPA, with at least a 50% improvement in annotation efficiency. Furthermore, the intersection-over-union (IoU) metric of our OBBOX is 3.6% higher than existing methods such as the “Segment Anything Model”. When applied in training, the model annotated with SPA shows a 4.7% higher mean average precision (mAP) compared to models using traditional annotation methods. These results confirm the technical advantages and practical impact of SPA in advancing small object detection in remote sensing.

Enhancements in image quality and block detection performance for Reinforced Soil-Retaining Walls under various illuminance conditions

To ensure continuous monitoring of reinforced soil-retaining walls (RSWs) even under low-illuminance conditions, such as during the night, it is imperative to evaluate the performance of deep learning-based detection. In this study, we constructed a laboratory RSW model and generated a dataset with varying illuminance levels to assess the impact of image enhancement and block detection. Various image enhancement methods were applied to improve image quality and evaluate their effect on deep learning. RGB optimization (RO) was proposed to optimize RGB intensity and compared with gamma correction, histogram equalization, and low-light image enhancement with illumination map estimation. RO demonstrated outstanding image enhancement performance, as evidenced by lightness order error, peak signal-to-noise ratio, and structural similarity index measure, ensuring high image quality. The trained RO model using Mask R-CNN exhibited excellent accuracy, recall, and F1 score, delivering remarkable detection performance under low illuminance conditions, resulting in a 7.44 % improvement in the F1 score. Image enhancement techniques that maintain similarity, such as lightness order error and structural similarity, across varying illuminance conditions contribute to enhancing the block detection performance of Mask R-CNNs.

Deep Learning-Based Detection of Oil Spills in Pakistan’s Exclusive Economic Zone from January 2017 to December 2023

Oil spillages on a sea’s or an ocean’s surface are a threat to marine and coastal ecosystems. They are mainly caused by ship accidents, illegal discharge of oil from ships during cleaning and oil seepage from natural reservoirs. Synthetic-Aperture Radar (SAR) has proved to be a useful tool for analyzing oil spills, because it operates in all-day, all-weather conditions. An oil spill can typically be seen as a dark stretch in SAR images and can often be detected through visual inspection. The major challenge is to differentiate oil spills from look-alikes, i.e., low-wind areas, algae blooms and grease ice, etc., that have a dark signature similar to that of an oil spill. It has been noted over time that oil spill events in Pakistan’s territorial waters often remain undetected until the oil reaches the coastal regions or it is located by concerned authorities during patrolling. A formal remote sensing-based operational framework for oil spills detection in Pakistan’s Exclusive Economic Zone (EEZ) in the Arabian Sea is urgently needed. In this paper, we report the use of an encoder–decoder-based convolutional neural network trained on an annotated dataset comprising selected oil spill events verified by the European Maritime Safety Agency (EMSA). The dataset encompasses multiple classes, viz., sea surface, oil spill, look-alikes, ships and land. We processed Sentinel-1 acquisitions over the EEZ from January 2017 to December 2023, and we thereby prepared a repository of SAR images for the aforementioned duration. This repository contained images that had been vetted by SAR experts, to trace and confirm oil spills. We tested the repository using the trained model, and, to our surprise, we detected 92 previously unreported oil spill events within those seven years. In 2020, our model detected 26 oil spills in the EEZ, which corresponds to the highest number of spills detected in a single year; whereas in 2023, our model detected 10 oil spill events. In terms of the total surface area covered by the spills, the worst year was 2021, with a cumulative 395 sq. km covered in oil or an oil-like substance. On the whole, these are alarming figures.