Detection Pipeline Research Articles

Continuous fake media detection: Adapting deepfake detectors to new generative techniques

Generative techniques continue to evolve at an impressively high rate, driven by the hype about these technologies. This rapid advancement severely limits the application of deepfake detectors, which, despite numerous efforts by the scientific community, struggle to achieve sufficiently robust performance against the ever-changing content. To address these limitations, in this paper, we propose an analysis of two continuous learning techniques on a Short and a Long sequence of fake media. Both sequences include a complex and heterogeneous range of deepfakes (generated images and videos) from GANs, computer graphics techniques, and unknown sources. Our experiments show that continual learning could be important in mitigating the need for generalizability. In fact, we show that, although with some limitations, continual learning methods help to maintain good performance across the entire training sequence. For these techniques to work in a sufficiently robust way, however, it is necessary that the tasks in the sequence share similarities. In fact, according to our experiments, the order and similarity of the tasks can affect the performance of the models over time. To address this problem, we show that it is possible to group tasks based on their similarity. This small measure allows for a significant improvement even in longer sequences. This result suggests that continual techniques can be combined with the most promising detection methods, allowing them to catch up with the latest generative techniques. In addition to this, we propose an overview of how this learning approach can be integrated into a deepfake detection pipeline for continuous integration and continuous deployment (CI/CD). This allows you to keep track of different funds, such as social networks, new generative tools, or third-party datasets, and through the integration of continuous learning, allows constant maintenance of the detectors.

Fishing event detection and species classification using computer vision and artificial intelligence for electronic monitoring

Fisheries regulations require detailed catch reporting on commercial fishing vessels. Vital components for the sustainable management of fish stocks include a robust estimate of the number of fish caught and the species composition. Catch recording is often done manually by human observers on fishing vessels. Human observers are costly, and consistent data streams can be subject to observer availability and the weather. On-vessel cameras (electronic monitoring, EM) are a growing alternative to human observers. However, on-land human auditors are required to review hundreds of hours of videos recorded during fishing trips that can last for weeks. In this paper, a framework is presented to automatically detect fish in EM videos, count the total fishing events, and classify the fish species. For this purpose, a deep learning and computer vision-based model is developed to efficiently detect fish and fishers onboard a vessel. Secondly, a vision-based tracking pipeline tracks the detected fish and counts the total fishing events in the videos. Thirdly, the extracted fishing events are classified through a deep learning-based fish species classifier, to provide the distribution of different fish species caught for a fishing trip. For our experiments, the datasets were prepared using the electronic monitoring data of multiple fishing trips of a fishing vessel. The videos were recorded on Australian longline vessels targeting tunas and billfish. For the fish detection task, video frames were extracted and labelled manually to provide a digital ground-truth. For the fish species classification task, hundreds of fish images of multiple species were cropped to provide a training dataset for the fish classifier. For the fish counting task, manual counts for the fishing events of individual fish species were generated for the test fishing trips. The developed fish and fisher detector achieves a mean Average Precision of 87.0 % for fish and 94.0 % for fishers on test video frames. The fishing event detection pipeline achieves an Average Precision of 81.0 % and an Average Recall of 74.5 % on test videos. The fish species classifier achieves an Accuracy (Top-1) of 91.11 % for the classification of cropped fish images and 89.05 % for the classification of extracted fishing events from the videos. Experimental results show that our proposed computer vision and artificial intelligence-based solution for video analysis has great potential to automate the auditing process from electronic monitoring footage and contribute to the sustainable management of fish stocks.

NEFELI: A deep-learning detection and tracking pipeline for enhancing autonomy in advanced air mobility

Efficient detection and accurate collision estimation for non-cooperative aerial vehicles are crucial for the realization of fully autonomous aircraft and Advanced Air Mobility (AAM). This paper introduces NEFELI, a machine learning software utilizing optical sensors to detect and track non-cooperative aerial vehicles. NEFELI's detector employs an enhanced YOLOv5-large model, strengthened with a sliced inference step to enhance detection capabilities for distant, diminutive objects. Furthermore, NEFELI introduces several innovations in its tracking component. A key advancement lies in the creation and utilization of the first large-scale re-identification (Re-ID) dataset of aerial objects. This dataset is used to train the deep learning appearance (Re-ID) model of the tracking module and integrates appearance information into the detection and tracking pipeline, resulting in more robust and reliable tracking performance. Moreover, the tracking model combines the deep learning appearance model with a Kalman Filter-based motion model to address the challenge of precisely tracking distant aerial objects. Notably, an extensive comparative analysis that was conducted showed that NEFELI outperforms state-of-the-art detection and tracking models in terms of Higher Order Tracking Accuracy (HOTA) metric, ID switches, and Association Accuracy (AssA) by a wide margin. A crucial aspect of this work is NEFELI's software architecture design, which enables efficient implementation on a low SWaP (Size, Weight, and Power) edge Graphic Processing Unit (GPU). To further showcase NEFELI's generalization capabilities and edge implementation performance, real-world flight experiments with small UAVs were carried out. The experimental results demonstrate NEFELI's ability to detect and track small UAVs at distances of up to 145 m in real-time speed of 6.7 fps.

LVI-PathNet: Segmentation-classification pipeline for detection of lymphovascular invasion in whole slide images of lung adenocarcinoma

Lymphovascular invasion (LVI) in lung cancer is a significant prognostic factor that influences treatment and outcomes, yet its reliable detection is challenging due to interobserver variability. This study aims to develop a deep learning model for LVI detection using whole slide images (WSIs) and evaluate its effectiveness within a pathologist's information system. Experienced pathologists annotated blood vessels and invading tumor cells in 162 WSIs of non-mucinous lung adenocarcinoma sourced from two external and one internal datasets. Two models were trained to segment vessels and identify images with LVI features. DeepLabV3+ model achieved an Intersection-over-Union of 0.8840 and an area under the receiver operating characteristic curve (AUC-ROC) of 0.9869 in vessel segmentation. For LVI classification, the ensemble model achieved a F1-score of 0.9683 and an AUC-ROC of 0.9987. The model demonstrated robustness and was unaffected by variations in staining and image quality. The pilot study showed that pathologists' evaluation time for LVI detecting decreased by an average of 16.95%, and by 21.5% in “hard cases”. The model facilitated consistent diagnostic assessments, suggesting potential for broader applications in detecting pathological changes in blood vessels and other lung pathologies.

Autonomous post‐disaster indoor navigation and survivor detection using low‐cost micro aerial vehicles

AbstractThis paper introduces an innovative autonomous survivor detection pipeline tailored for low‐cost micro aerial vehicles (MAVs) operating in post‐disaster indoor environments. This consists of three main components: (1) a novel pipeline for survivor geotagging, which includes autonomous navigation, mapping, and detection of survivors using thermal images; (2) a navigation strategy to ensure complete thermal scanning coverage for survivor detection using low‐cost commercial grade thermal camera; and (3) robust and accurate survivor detection using YOLOv8x and thermal imaging. To demonstrate the effectiveness of the proposed framework, first, the autonomous navigation algorithm is simulated in Robotic Operating System (ROS) and experimentally validated under different layouts. Second, the YOLOv8x algorithm is pretrained and achieves high accuracy. Finally, a real‐world implementation was conducted with partially covered survivors in a simulated post‐disaster environment. The results demonstrated the proposed pipeline can accurately map the layout of the environment and identify all survivors. This study demonstrates that affordable MAVs with basic thermal cameras can be effectively used to geotag survivors to support rescue missions during post‐disaster events.

Detection and localization of the highly active FRB 20240114A with MeerKAT

ABSTRACT We report observations of the highly active FRB 20240114A with MeerKAT using the ultra-high frequency (UHF; $544\!-\!1088$ MHz) and L band ($856\!-\!1712$ MHz) receivers. A total of 62 bursts were detected in coherent tied-array beams using the MeerTRAP real-time transient detection pipeline. We measure a structure-optimizing dispersion measure of $527.65\pm 0.01\, \text{pc}\, \text{cm}^{-3}$ using the brightest burst in the sample. We find the bursts of FRB 20240114A are generally detected in part of the broad-band of MeerKAT, $\sim 40~{{\ \rm per\ cent}}$ in the UHF and $\sim 30~{{\ \rm per\ cent}}$ in the L band, indicating the band limited nature. We analyse the fluence distribution of the 44 bursts detected at UHF, constraining the fluence completeness limit to $\sim 1$ Jy ms, above which the cumulative burst rate follows a power law $R (\gt F)\propto (F/1\, \text{Jy}\, \text{ms})^\gamma$ with $\gamma =-1.8\pm 0.2$. Using channelized telescope data captured in our transient buffer we localize FRB 20240114A in the image domain to RA = 21h27m39.86s, Dec. = +04d19m45.01s with an uncertainty of 1.4 arcsec. This localization allows us to confidently identify the host galaxy of FRB 20240114A. Also using the transient buffer data, we perform a polarimetric study and demonstrate that most of the bursts have $\sim 100~{{\ \rm per\ cent}}$ linear polarization fractions and up to $\sim 20~{{\ \rm per\ cent}}$ circular polarization fractions. Finally, we predict the flux density of a potential persistent radio source (PRS) associated with FRB 20240114A is $\backsimeq [0.6\!-\!60]\, \mu \text{Jy}$ based on the simple relation between the luminosity of the PRS and the rotation measure arising from the FRB local environment.

DART: An automated end-to-end object detection pipeline with data Diversification, open-vocabulary bounding box Annotation, pseudo-label Review, and model Training

Accurate real-time object detection is vital across numerous industrial applications, from safety monitoring to quality control. Traditional approaches, however, are hindered by arduous manual annotation and data collection, struggling to adapt to ever-changing environments and novel target objects. To address these limitations, this paper presents DART, an innovative automated end-to-end pipeline that revolutionizes object detection workflows from data collection to model evaluation. It eliminates the need for laborious human labeling and extensive data collection while achieving outstanding accuracy across diverse scenarios. DART encompasses four key stages: (1) Data Diversification using subject-driven image generation (DreamBooth with SDXL), (2) Annotation via open-vocabulary object detection (Grounding DINO) to generate bounding box and class labels, (3) Review of generated images and pseudo-labels by large multimodal models (InternVL-1.5 and GPT-4o) to guarantee credibility, and (4) Training of real-time object detectors (YOLOv8 and YOLOv10) using the verified data. We apply DART to a self-collected dataset of construction machines named Liebherr Product, which contains over 15K high-quality images across 23 categories. The current instantiation of DART significantly increases average precision (AP) from 0.064 to 0.832. Its modular design ensures easy exchangeability and extensibility, allowing for future algorithm upgrades, seamless integration of new object categories, and adaptability to customized environments without manual labeling and additional data collection. The code and dataset are released at https://github.com/chen-xin-94/DART.

Enhanced human motion detection with hybrid RDA-WOA-based RNN and multiple hypothesis tracking for occlusion handling

Human motion detection in complex scenarios poses challenges due to occlusions. This paper presents an integrated approach for accurate human motion detections by combining Adapted Canny Edge detection as a preprocessing step, backbone-modified Mask R-CNN for precise segmentation, Hybrid RDA-WOA-based RNN as a classification, and a Multiple-hypothesis model for effective occlusion handling. Adapted Canny Edge detection is utilized as an initial preprocessing step to highlight significant edges in the input image. The resulting edge map enhances object boundaries and highlights structural features, simplifying subsequent processing steps. The improved image is then passed through backbone-modified Mask R-CNN for the pixel-level segmentation of humans. Backbone-modified Mask R-CNN along with IoU, Euclidean Distance, and Z-Score recognizes moving objects in complex scenes exactly. After recognizing moving objects, the optimized Hybrid RDA-WOA-based RNN classifies humans. To handle the self-occlusion, Multiple Hypothesis Tracking (MHT) is used. Real-world situations frequently include occlusions where humans can be partially or completely hidden by objects. The proposed approach integrates a Multiple-hypothesis model into the detection pipeline to address this challenge. Moreover, the proposed human motion detection approach includes an optimized Hybrid RDA-WOA-based RNN trained with 2D representations of 3D skeletal motion. The proposed work was evaluated using the IXMAS, KTH, Weizmann, NTU RGB + D, and UCF101 Datasets. It achieved an accuracy of 98% on the IXMAS, KTH, Weizmann, and UCF101 Datasets and 97.1% on the NTU RGB + D Dataset. The simulation results unveil the superiority of the proposed methodology over the existing approaches.

CircCNNs, a convolutional neural network framework to better understand the biogenesis of exonic circRNAs

Circular RNAs (circRNAs) as biomarkers for cancer detection have been extensively explored, however, the biogenesis mechanism is still elusive. In contrast to linear splicing (LS) involved in linear transcript formation, the so-called back splicing (BS) process has been proposed to explain circRNA formation. To investigate the potential mechanism of BS via the machine learning approach, we curated a high-quality BS and LS exon pairs dataset with evidence-based stringent filtering. Two convolutional neural networks (CNN) base models with different structures for processing splicing junction sequences including motif extraction were created and compared after extensive hyperparameter tuning. In contrast to the previous study, we are able to identify motifs corresponding to well-established BS-associated genes such as MBNL1, QKI, and ESPR2. Importantly, despite prevalent high false positive rates in existing circRNA detection pipelines and databases, our base models demonstrated a notable high specificity (greater than 90%). To further improve the model performance, a novo fast numerical method was proposed and implemented to calculate the reverse complementary matches (RCMs) crossing two flanking regions and within each flanking region of exon pairs. Our CircCNNs framework that incorporated RCM information into the optimal base models further reduced the false positive rates leading to 88% prediction accuracy.

VolcanoSV enables accurate and robust structural variant calling in diploid genomes from single-molecule long read sequencing

Structural variants (SVs) significantly contribute to human genome diversity and play a crucial role in precision medicine. Although advancements in single-molecule long-read sequencing offer a groundbreaking resource for SV detection, identifying SV breakpoints and sequences accurately and robustly remains challenging. We introduce VolcanoSV, an innovative hybrid SV detection pipeline that utilizes both a reference genome and local de novo assembly to generate a phased diploid assembly. VolcanoSV uses phased SNPs and unique k-mer similarity analysis, enabling precise haplotype-resolved SV discovery. VolcanoSV is adept at constructing comprehensive genetic maps encompassing SNPs, small indels, and all types of SVs, making it well-suited for human genomics studies. Our extensive experiments demonstrate that VolcanoSV surpasses state-of-the-art assembly-based tools in the detection of insertion and deletion SVs, exhibiting superior recall, precision, F1 scores, and genotype accuracy across a diverse range of datasets, including low-coverage (10x) datasets. VolcanoSV outperforms assembly-based tools in the identification of complex SVs, including translocations, duplications, and inversions, in both simulated and real cancer data. Moreover, VolcanoSV is robust to various evaluation parameters and accurately identifies breakpoints and SV sequences.

Radiomics signature for automatic hydronephrosis detection in unenhanced Low-Dose CT

PurposeTo investigate the diagnostic performance of an automatic pipeline for detection of hydronephrosis on kidney's parenchyma on unenhanced low-dose CT of the abdomen. MethodsThis retrospective study included 95 patients with confirmed unilateral hydronephrosis in an unenhanced low-dose CT of the abdomen. Data were split into training (n = 67) and test (n = 28) cohorts. Both kidneys for each case were included in further analyses, whereas the kidney without hydronephrosis was used as control. Using the training cohort, we developed a pipeline consisting of a deep-learning model for automatic segmentation (a Convolutional Neural Network based on nnU-Net architecture) of the kidney's parenchyma and a radiomics classifier to detect hydronephrosis. The models were assessed using standard classification metrics, such as area under the ROC curve (AUC), sensitivity and specificity, as well as semantic segmentation metrics, including Dice coefficient and Jaccard index. ResultsUsing manual segmentation of the kidney’s parenchyma, hydronephrosis can be detected with an AUC of 0.84, a sensitivity of 75% and a specificity of 82%, a PPV of 81% and a NPV of 77%. Automatic kidney segmentation achieved a mean Dice score of 0.87 and 0.91 for the right and left kidney, respectively. Additionally, automatic segmentation achieved an AUC of 0.83, a sensitivity of 86%, specificity of 64%, PPV of 71%, and NPV of 82%. ConclusionOur proposed radiomics signature using automatic kidney's parenchyma segmentation allows for accurate hydronephrosis detection on unenhanced low-dose CT scans of the abdomen independently of widened renal pelvis. This method could be used in clinical routine to highlight hydronephrosis to radiologists as well as clinicians, especially in patients with concurrent parapelvic cysts and might reduce time and costs associated with diagnosing hydronephrosis.

Impact of Perception Errors in Vision-Based Detection and Tracking Pipelines on Pedestrian Trajectory Prediction in Autonomous Driving Systems.

Pedestrian trajectory prediction is crucial for developing collision avoidance algorithms in autonomous driving systems, aiming to predict the future movement of the detected pedestrians based on their past trajectories. The traditional methods for pedestrian trajectory prediction involve a sequence of tasks, including detection and tracking to gather the historical movement of the observed pedestrians. Consequently, the accuracy of trajectory prediction heavily relies on the accuracy of the detection and tracking models, making it susceptible to their performance. The prior research in trajectory prediction has mainly assessed the model performance using public datasets, which often overlook the errors originating from detection and tracking models. This oversight fails to capture the real-world scenario of inevitable detection and tracking inaccuracies. In this study, we investigate the cumulative effect of errors within integrated detection, tracking, and trajectory prediction pipelines. Through empirical analysis, we examine the errors introduced at each stage of the pipeline and assess their collective impact on the trajectory prediction accuracy. We evaluate these models across various custom datasets collected in Taiwan to provide a comprehensive assessment. Our analysis of the results derived from these integrated pipelines illuminates the significant influence of detection and tracking errors on downstream tasks, such as trajectory prediction and distance estimation.

Improved allele-specific single-cell copy number estimation in low-coverage DNA-sequencing.

Advances in whole-genome single-cell DNA sequencing (scDNA-seq) have led to the development of numerous methods for detecting copy number aberrations (CNAs), a key driver of genetic heterogeneity in cancer. While most of these methods are limited to the inference of total copy number, some recent approaches now infer allele-specific CNAs using innovative techniques for estimating allele-frequencies in low coverage scDNA-seq data. However, these existing allele-specific methods are limited in their segmentation strategies, a crucial step in the CNA detection pipeline. We present SEACON (Single-cell Estimation of Allele-specific COpy Numbers), an allele-specific copy number profiler for scDNA-seq data. SEACON uses a Gaussian Mixture Model to identify latent copy number states and breakpoints between contiguous segments across cells, filters the segments for high-quality breakpoints using an ensemble technique, and adopts several strategies for tolerating noisy read-depth and allele frequency measurements. Using a wide array of both real and simulated datasets, we show that SEACON derives accurate copy numbers and surpasses existing approaches under numerous experimental conditions, and identify its strengths and weaknesses. SEACON is implemented in Python and is freely available open-source from https://github.com/NabaviLab/SEACON and https://doi.org/10.5281/zenodo.12727008.

Constraining the Selection-corrected Luminosity Function and Total Pulse Count for Radio Transients

Studying transient phenomena, such as individual pulses from pulsars, has garnered considerable attention in the era of astronomical big data. Of specific interest to this study are rotating radio transients (RRATs), nulling, and intermittent pulsars. This study introduces a new algorithm named LuNfit, tailored to correct the selection biases originating from the telescope and detection pipelines. Ultimately LuNfit estimates the intrinsic luminosity distribution and nulling fraction of the single pulses emitted by pulsars. LuNfit relies on Bayesian nested sampling so that the parameter space can be fully explored. Bayesian nested sampling also provides the additional benefit of simplifying model comparisons through the Bayes ratio. The robustness of LuNfit is shown through simulations and applying LuNfit onto pulsars with known nulling fractions. LuNfit is then applied to three RRATs, J0012+5431, J1538+1523, and J2355+1523, extracting their intrinsic luminosity distribution and burst rates. We find that their nulling fractions are 0.4(2), 0.749(5), and 0.995(2), respectively. We further find that a log-normal distribution likely describes the single pulse luminosity distribution of J0012+5431 and J1538+1523, while the Bayes ratio for J2355+1523 slightly favors an exponential distribution. We show the conventional method of correcting selection effects by “scaling up” the missed fraction of radio transients can be unreliable when the mean luminosity of the source is faint relative to the telescope sensitivity. Finally, we discuss the limitations of the current implementation of LuNfit while also delving into potential enhancements that would enable LuNfit to be applied to sources with complex pulse morphologies.

Evidence of Grapevine enamovirus 1 Infecting Grapevine (Vitis vinifera L.) in Canada.

Grapevine enamovirus 1 (GEV1) belongs to the genus Enamovirus, in the family Solemoviridae. It has been reported from several countries infecting grapevines including Brazil (Silva et al. 2017), China (Ren et al. 2021) and France (Hily et al. 2022). To assess the prevalence and diversity of economically important grapevine viruses in nine Canadian vineyards, total RNA and double-stranded RNA (dsRNA) (Fall et al. 2020) were extracted from 30 and 100 composite samples respectively, with each consisting of five vines of the same cultivars. The cultivars included in this study are Frontenac noir (n=34), Vidal (n=32), Marquette (n=33), Riesling (n=31), and Pinot noir (n=31). The total RNA and dsRNA samples were subsequently multiplexed and diagnosed by high-throughput sequencing (HTS) on NovaSeq (600 S4 PE100) and MiSeq (2 × 250 cycle PE) respectively. From NovaSeq and MiSeq sequencing, an average of 410,000 to 1.3 million reads/sample were obtained, respectively, with mapped viral reads representing 10.92% to 12.48% of the total reads. After sequence quality was verified using Trimmomatic v.0.40 (Bolger et al. 2014), the clean sequences were screened against all possible viruses in the databases using the Virtool (Rott et al. 2017) and VirFind virus detection pipelines (Ho and Tzanetakis 2014). GEV1 was detected in clean sequences from two, three, and two leaf samples of cultivars 'Marquette' 'Riesling' and 'Frontenac noir' respectively. Six of the seven HTS-assembled GEV1 genomes were partial, ranging from 4,523 to 6,000 nucleotide (nt) with genome coverage varying from 71% to 89%. Only one 6,314 nt long assembled contig (Accession No. OR021829), represented a nearly complete genome, being only 53 and 3 nt shorter than Sd-CG (MT536978) at 5' and 3' untranslated regions (UTR), respectively. Isolate 3- Riesling-CAN (OR021829) shares 90.56 to 94.19% nt identities with several GEV1isolates at 96-99% of query coverage. Phylogenetically, OR021829 is closer to GEV1 isolates from France and China (Figure S1). To validate the HTS results, the developed primer pair SetF and Set1R (Silva et al., 2017) was used for RT-PCR detection. The amplicons from all seven HTS-positive samples were sequenced using Sanger sequencing, confirming the presence of GEV-1 in three studied grape cultivars in Canadian vineyards. Symptoms associated with the specific GEV1-infected vines could not be explained as composite samples were used. Each of the combined samples HTS library also tested positive for at least one of the known grape virus/viroids, namely grapevine leafroll associated-virus -3, grapevine pinot gris virus, grapevine rupestris stem pitting-associated virus, Marafivirus syrahense grapevine Syrah virus-1 and hop stunt viroid. To our knowledge, this is the first report of GEV1 being detected in grapevines in Canada, or in any North American vineyard. GEV1 is a relatively new virus, and its biology remains largely unknown. Based on this sequence new GEV1 primers can be developed to know the genetic variability among GEV-1 and improve the detection of this virus in vineyards.

OPTICAL DEEP LEARNING LANDMINE DETECTION BASED ON LIMITED DATASET OF AERIAL IMAGERY

Landmine detection is one of the most innovative applications of unmanned aerial vehicles that became possible due to rapid development of both aerial vehicles equipped by different optical cameras and sensors using different physical principles, and object classification and detection methods, including machine learning and especially deep learning. Optical detection is an essential part of the overall landmine detection process that can be performed either separately or in combination with data processing from other types of cameras or sensors. The development of deep convolutional neural networks has dramatically changed the landscape of optical detection by making them de-facto choice number one for the majority of object classification, detection and segmentation tasks. However, the deterrent factor in the case of landmine detection is limited availability of appropriate data for training that different researchers try to overcome in different ways. The assessment of necessary amount of training data for any particular object detection problem still remains an experimental task. Despite several years of development in this area, still there is a shortage of research based on real landmine imagery obtained from unmanned aerial vehicles, so currently any public effort in this direction is valuable and works as an inspiration for new researchers. This paper describes such a study, namely its first iteration in which popular open-source tools are used to build detection pipeline and estimation of their efficiency is done using limited amount of data. It is shown that the problem of limited amount of training data can be effectively overcome by data augmentation and iterational process of training optical landmine detector is demonstrated. The effectiveness of open-source tools and libraries for neural networks training, object detection and dataset preparation is also demonstrated.

Lane Detection, Segmentation, Pothole Detection and Traffic Sign Recognition for ADAS

Lane detection, a critical aspect of advanced driver assistance systems (ADAS) and autonomous vehicles, is addressed in this work, combining classical computer vision methods with deep learning techniques. The proposed solution, utilizing a UNet-based model trained with TensorFlow and Keras, enhances vehicle perception for intelligent transportation systems. Integrated functionalities for pothole detection and traffic sign detection further contribute to safety and efficiency, enabling vehicles to identify road hazards and comply with regulations. The system encompasses data preprocessing, model training, and real-time video analysis, while a classical lane detection pipeline using OpenCV showcases various stages such as grayscale conversion, Gaussian blur, Canny edge detection, masking, Hough transform, and lane overlay. This comprehensive approach supports road safety, traffic management, and transportation efficiency initiatives, making significant strides in intelligent transportation systems and urban planning.

Application of multiple mosaic callers improves post-zygotic mutation detection from exome sequencing data

PurposeThe gold standard for identification of post-zygotic variants (PZVs) is droplet digital polymerase chain reaction or high-depth sequencing across multiple tissues types. These approaches are yet to be systematically implemented for monogenic disorders. We developed PZV detection pipelines for correct classification of de novo variants. MethodOur pipelines detect PZV in parents (gonosomal mosaicism [pGoM]) and children (somatic mosaicism, “M3”). We applied them to research exome sequencing (ES) data from the Australian Cerebral Palsy Biobank (n = 145 trios) and Simons Simplex Collection (n = 405 families). Candidate mosaic variants were validated using deep amplicon sequencing or droplet digital polymerase chain reaction. Results69.2% (M3trio), 63.9% (M3single), and 92.7% (pGoM) of detected variants were validated, with 48.6%, 56.7%, and 26.2% of variants, respectively, meeting strict criteria for mosaicism. In the Australian Cerebral Palsy Biobank, 16.6% of probands and 20.7% of parents had at least 1 true-positive somatic or pGoM variant, respectively. A large proportion of PZVs detected in Simons Simplex Collection parents (79.8%) and child (94.5%) were not previously reported. We reclassified 3.7% to 8.0% of germline de novo variants as mosaic. ConclusionMany PZVs were incorrectly classified as germline variants or missed by previous approaches. Systematic application of our pipelines could increase genetic diagnostic rate, improve estimates of recurrence risk in families, and benefit novel disease gene identification.

Automatized detection of uniparental disomies in a large cohort

Uniparental disomy (UPD) is the inheritance of both homologues of a chromosome from only one parent. The detection of UPDs in sequencing data is not well established and a common gap in genetic diagnostics. We applied our in-house UPD detection pipeline to evaluate a cohort of 9212 samples, including multigene panels as well as exome sequencing data in a single, duo or trio constellation. We used the results to inform the design of our publicly available web app altAFplotter. UPDs categorized as heterodisomy, whole chromosome or segmental isodisomy were identified and validated with microsatellites, multiplex ligation-dependent probe amplification as well as Sanger sequencing. We detected 14 previously undiagnosed UPDs including nine isodisomies, four segmental isodisomies as well as one heterodisomy on chromosome 22. We characterized eight findings as potentially causative through homozygous pathogenic variants or imprinting disorders. Overall, our study demonstrates the utility of our UPD detection pipeline with our web app, altAFplotter, to reliably identify UPDs. This not only increases the diagnostic yield of cases with growth and metabolic disturbances, as well as developmental delay, but also enhances the understanding of UPDs that may be relevant for recurrence risks and genetic counseling.

Comprehensive and deep evaluation of structural variation detection pipelines with third-generation sequencing data

BackgroundStructural variation (SV) detection methods using third-generation sequencing data are widely employed, yet accurately detecting SVs remains challenging. Different methods often yield inconsistent results for certain SV types, complicating tool selection and revealing biases in detection.ResultsThis study comprehensively evaluates 53 SV detection pipelines using simulated and real data from PacBio (CLR: Continuous Long Read, CCS: Circular Consensus Sequencing) and Nanopore (ONT) platforms. We assess their performance in detecting various sizes and types of SVs, breakpoint biases, and genotyping accuracy with various sequencing depths. Notably, pipelines such as Minimap2-cuteSV2, NGMLR-SVIM, PBMM2-pbsv, Winnowmap-Sniffles2, and Winnowmap-SVision exhibit comparatively higher recall and precision. Our findings also show that combining multiple pipelines with the same aligner, like pbmm2 or winnowmap, can significantly enhance performance. The individual pipelines’ detailed ranking and performance metrics can be viewed in a dynamic table: http://pmglab.top/SVPipelinesRanking.ConclusionsThis study comprehensively characterizes the strengths and weaknesses of numerous pipelines, providing valuable insights that can improve SV detection in third-generation sequencing data and inform SV annotation and function prediction.