Multi-task Convolutional Neural Network Research Articles

Direct reconstruction for simultaneous dual-tracer PET imaging based on multi-task learning

BackgroundSimultaneous dual-tracer positron emission tomography (PET) imaging can observe two molecular targets in a single scan, which is conducive to disease diagnosis and tracking. Since the signals emitted by different tracers are the same, it is crucial to separate each single tracer from the mixed signals. The current study proposed a novel deep learning-based method to reconstruct single-tracer activity distributions from the dual-tracer sinogram.MethodsWe proposed the Multi-task CNN, a three-dimensional convolutional neural network (CNN) based on a framework of multi-task learning. One common encoder extracted features from the dual-tracer dynamic sinogram, followed by two distinct and parallel decoders which reconstructed the single-tracer dynamic images of two tracers separately. The model was evaluated by mean squared error (MSE), multiscale structural similarity (MS-SSIM) index and peak signal-to-noise ratio (PSNR) on simulated data and real animal data, and compared to the filtered back-projection method based on deep learning (FBP-CNN).ResultsIn the simulation experiments, the Multi-task CNN reconstructed single-tracer images with lower MSE, higher MS-SSIM and PSNR than FBP-CNN, and was more robust to the changes in individual difference, tracer combination and scanning protocol. In the experiment of rats with an orthotopic xenograft glioma model, the Multi-task CNN reconstructions also showed higher qualities than FBP-CNN reconstructions.ConclusionsThe proposed Multi-task CNN could effectively reconstruct the dynamic activity images of two single tracers from the dual-tracer dynamic sinogram, which was potential in the direct reconstruction for real simultaneous dual-tracer PET imaging data in future.

A multi-task convolutional neural network for classification and segmentation of chronic venous disorders

Chronic Venous Disorders (CVD) of the lower limbs are one of the most prevalent medical conditions, affecting 35% of adults in Europe and North America. Due to the exponential growth of the aging population and the worsening of CVD with age, it is expected that the healthcare costs and the resources needed for the treatment of CVD will increase in the coming years. The early diagnosis of CVD is fundamental in treatment planning, while the monitoring of its treatment is fundamental to assess a patient’s condition and quantify the evolution of CVD. However, correct diagnosis relies on a qualitative approach through visual recognition of the various venous disorders, being time-consuming and highly dependent on the physician’s expertise. In this paper, we propose a novel automatic strategy for the joint segmentation and classification of CVDs. The strategy relies on a multi-task deep learning network, denominated VENet, that simultaneously solves segmentation and classification tasks, exploiting the information of both tasks to increase learning efficiency, ultimately improving their performance. The proposed method was compared against state-of-the-art strategies in a dataset of 1376 CVD images. Experiments showed that the VENet achieved a classification performance of 96.4%, 96.4%, and 97.2% for accuracy, precision, and recall, respectively, and a segmentation performance of 75.4%, 76.7.0%, 76.7% for the Dice coefficient, precision, and recall, respectively. The joint formulation increased the robustness of both tasks when compared to the conventional classification or segmentation strategies, proving its added value, mainly for the segmentation of small lesions.

Joint CD, DGD, and PDL Estimation Enabled by FrFT based Time-Frequency Reconstruction

A high-precision optical performance monitoring (OPM) scheme is proposed to achieve chromatic dispersion estimation (CDE), differential group delay (DGD) estimation, and polarization-dependent loss (PDL) estimation simultaneously for the long haul coherent optical fiber communication system. A time-frequency analysis method based on the fractional Fourier transformation (FrFT) is applied to the optical signal under different channel impairments to reconstruct the two-dimensional distribution images. Multi-task convolutional neural network (MT-CNN) is then adopted to extract the corresponding features of impairments and establish a solid relationship between the images and impairment quantities, thus can jointly estimate CD, DGD, and PDL. We validate in simulation the proposed estimation using 50GBaud PDM-16QAM, the mean absolute error (MAE) for CD, DGD, and PDL estimation is 114ps/nm, 0.26ps, and 0.072dB, and the monitoring window ranges from 1600∼48000ps/nm, 4∼100ps, and 0∼2dB, respectively. Simulation verification indicates that the proposed estimation method achieves high precision and robustness to ASE noise and fiber nonlinearity of self-phase modulation (SPM).

Self-supervised EEG emotion recognition models based on CNN.

Emotion plays crucial roles in human life. Recently, emotion classification from electroencephalogram (EEG) signal has attracted attention by researchers due to the rapid development of brain computer interface (BCI) techniques and machine learning algorithms. However, recent studies on emotion classification show resource utilization because they use the fully-supervised learning methods. Therefore, in this study, we applied the self-supervised learning methods to improve the efficiency of resources usage. We employed a self-supervised approach to train deep multi-task convolutional neural network (CNN) for EEG-based emotion classification. First, six signal transformations were performed on unlabeled EEG data to construct the pretext task. Second, a multi-task CNN was used to perform signal transformation recognition on the transformed signals together with the original signals. After the signal transformation recognition network was trained, the convolutional layer network was frozen and the fully connected layer was reconstructed as emotion recognition network. Finally, the EEG data with affective labels were used to train the emotion recognition network to clarify the emotion. In this paper, we conduct extensive experiments from the data scaling perspective using the SEED, DEAP affective dataset. Results showed that the self-supervised learning methods can learn the internal representation of data and save computation time compared to the fully-supervised learning methods. In conclusion, our study suggests that the self-supervised machine learning model can improve the performance of emotion classification compared to the conventional fully supervised model.

MRI-based classification of IDH mutation and 1p/19q codeletion status of gliomas using a 2.5D hybrid multi-task convolutional neural network.

IDH mutation and 1p/19q codeletion status are important prognostic markers for glioma that are currently determined using invasive procedures. Our goal was to develop artificial intelligence-based methods to noninvasively determine molecular alterations from MRI. Pre-operative MRI scans of 2648 glioma patients were collected from Washington University School of Medicine (WUSM; n = 835) and publicly available Brain Tumor Segmentation (BraTS; n = 378), LGG 1p/19q (n = 159), Ivy Glioblastoma Atlas Project (Ivy GAP; n = 41), The Cancer Genome Atlas (TCGA; n = 461), and the Erasmus Glioma Database (EGD; n = 774) datasets. A 2.5D hybrid convolutional neural network was proposed to simultaneously localize glioma and classify its molecular status by leveraging MRI imaging features and prior knowledge features from clinical records and tumor location. The models were trained on 223 and 348 cases for IDH and 1p/19q tasks, respectively, and tested on one internal (TCGA) and two external (WUSM and EGD) test sets. For IDH, the best-performing model achieved areas under the receiver operating characteristic (AUROC) of 0.925, 0.874, 0.933 and areas under the precision-recall curves (AUPRC) of 0.899, 0.702, 0.853 on the internal, WUSM, and EGD test sets, respectively. For 1p/19q, the best model achieved AUROCs of 0.782, 0.754, 0.842, and AUPRCs of 0.588, 0.713, 0.782, on those three data-splits, respectively. The high accuracy of the model on unseen data showcases its generalization capabilities and suggests its potential to perform "virtual biopsy" for tailoring treatment planning and overall clinical management of gliomas.

A multi-task learning approach for chemical process abnormity locations and fault classifications

The fault classification of chemical processes based on supervised learning generally demands for enough labeled fault samples, which represents a big challenge for practical engineering issues. In the industrial field, when capturing an appearance of process abnormities, human operators usually take account of correlations of relevant variables so as to locate process variables associated with the abnormity before further classifying or diagnosing the corresponding faults. In this context, accurate abnormity localizations can be an effective help for classifying the fault type. Nowadays, deep learning based fault classifications have been extensively studied and widely circulated in the literature. However, the existing approaches rarely concern the supporting role of abnormity locations for fault classifications. Motivated by this observation, this paper proposes a multi-task learning (MTL) based convolutional neural network (CNN) model for concurrent abnormity locations and fault classifications for chemical processes. Specifically, fault samples are labeled with both abnormal variable locations and fault categories before constructing a multi-task learning CNN classifier performing an auxiliary task, abnormity locations, and a primary task, fault classifications. In addition, dynamic task weights are assigned to facilitate task contributions. This method can enjoy effectively reducing the necessary amounts of labeled fault samples in achieving satisfactory fault classification performances. Experiments are carried out on the TE process simulation platform and an industrial coal gasification plant along with comparisons with conventional fault classification methods, verifying the effectiveness of the proposed method.

Advances in multi- and hyperspectral remote sensing of mangrove species: A synthesis and study case on airborne and multisource spaceborne imagery

This study summarizes the advances in mangrove species mapping based on multispectral and hyperspectral imagery achieved over the last decade. The influence of species diversity and sensor specifications on the performances of various classification approaches are discussed. Based on the limitations of previous approaches, we propose a novel framework to map mangrove species at medium, high, and very-high spatial resolution using multispectral and hyperspectral images. The framework relies on a multitask convolutional neural network to achieve accurate classification at pixel and object (i.e. individual tree crown) level. It was successfully applied to the most comprehensive dataset of optical imagery ever collected over a mangrove forest, achieving accurate mapping of species on airborne (OA = 95 %, Kappa = 0.93) and spaceborne imagery (OA and Kappa up to 97 % and 0.95, respectively), including the newly-operating hyperspectral DESIS and PRISMA instruments. This study brings new insights into the role of spatial and spectral resolutions in mangrove species classification, particularly the importance of short-wave infrared bands. Multispectral imagery performs well at very-high to high resolutions (up to 10–20 m) but suffers from the lack of spectral information at medium resolution (30 m). Hyperspectral imagery provided the best results at sub-metric resolution (0.5 m) and satisfactory results at 30-m resolution (OA ≥ 85), and could benefit from implementing spectral unmixing in the latter case. Given the increasing attention accorded to mangroves, remote sensing will undoubtedly become an essential tool to monitor the diversity of these endangered ecosystems in the future. In that respect, efforts should concentrate on evaluating the capabilities of new and upcoming instruments and multisource data combination to improve mangrove species mapping and, more broadly, to meet conservation goals.

Interpretation of multi-task clearance models from molecular images supported by experimental design

Recent methodological advances in deep learning (DL) architectures have not only improved the performance of predictive models but also enhanced their interpretability potential, thus considerably increasing their transparency. In the context of medicinal chemistry, the potential to not only accurately predict molecular properties, but also chemically interpret them, would be strongly preferred. Previously, we developed accurate multi-task convolutional neural network (CNN) and graph convolutional neural network (GCNN) models to predict a set of diverse intrinsic metabolic clearance parameters from image- and graph-based molecular representations, respectively. Herein, we introduce several model interpretability frameworks to answer whether the model explanations obtained from CNN and GCNN multi-task clearance models could be applied to predict chemical transformations associated with experimentally confirmed metabolic products. We show a strong correlation between the CNN pixel intensities and corresponding clearance predictions, as well as their robustness to different molecular orientations. Using actual case examples, we demonstrate that both CNN and GCNN interpretations frequently complement each other, suggesting their high potential for combined use in guiding medicinal chemistry design.

Multi-task learning on the edge for effective gender, age, ethnicity and emotion recognition

More and more real-world applications, especially in cognitive robotics, require the running of different computer vision algorithms in parallel on board of embedded devices with limited GPU and memory resources. Multi-task learning, namely the usage of the same model to perform multiple classification and/or regression tasks by learning a shared low level representation, revealed to be a valid solution to reduce computation and required memory space while preserving the accuracy. In this paper, we propose a solution for real-time user profiling based on a multi-task convolutional neural network (CNN) for gender, age, ethnicity and emotion recognition from face images. To find the best trade-off between accuracy and processing time, we evaluate three different architectures, specifically designed for the purpose, and backbones, based on MobileNet, ResNet and SENet, which include convolutional layers, residual blocks and attention modules that already demonstrated great potential in face analysis. We trained the multi-task neural network with a custom learning procedure, which solves the problems of missing labels, dataset imbalance and loss function imbalance through label masking, batch balancing and a custom weighted loss function; there are no other multi-task neural networks for face analysis that address all these challenges simultaneously.. The proposed solution demonstrated its effectiveness in the comparison with the corresponding single-task CNNs in terms of accuracy, processing time and memory space; in fact, the multi-task CNNs achieved a processing speed-up between 2.5 and 4 times and a reduction of the memory space between 2 and 4 times, while preserving the accuracy. Moreover, the useful insights that arise from the experiments allow to choose a solution for face analysis easily integrable into real applications on smart cameras and embedded systems and most suited for the specific application constraints in terms of computational resources.

Semantic-aware multi-task learning for image aesthetic quality assessment

In recent years, image aesthetic quality assessment has attracted considerable attention due to the massive growth of digital images in social platforms and the Internet. However, automatically assessing aesthetic quality of an image is a challenging task, because image aesthetic is affected by various factors, and the criteria for judging the aesthetic of images with diverse semantic information are different. To this end, a Semantic-Aware Multi-task convolution neural network (SAM-CNN) for evaluating image aesthetic quality is proposed in this paper. The network can fuse intermediate features of different layers at different scales in CNN to obtain a more comprehensive and accurate aesthetic expression, under the joint supervision of image aesthetic quality assessment task and semantic classification task in a multi-task learning manner. Besides, by applying the attention mechanism, semantic information with a large receptive field extracted from deep layers is utilised to guide the network to focus on the key parts of features to be fused, to improve the effectiveness of feature fusion. Experimental results on the AVA dataset and Photo.net dataset demonstrate the effectiveness and superiority of the proposed SAM-CNN.

D[formula omitted]-Net: Integrated multi-task convolutional neural network for water surface deblurring, dehazing and object detection

Visual perception is one of the key technologies for smart ships to achieve intelligent navigation in various complex water scenes, which adopts the object detection algorithm based on visible image to detect obstacles. Affected by the rolling and high-speed motion of ships and bad weather, motion blur and haze appear in the images of vision sensor, which seriously interfere with the obstacle detection. To ensure the navigation safety of smart ships, we propose the integrated multi-task deblurring, dehazing and object detection convolutional neural network (D3-Net). Firstly, the proposed method is able to complete deblurring, dehazing and object detection within a single network, which is more suitable to be applied in intelligent navigation. Secondly, different from traditional methods that reconstruct enhanced images first and then perform object detection, D3-Net combines image reconstruction and detection feature extraction to realize the integration of image enhancement and detection procedure. Finally, to improve the deblurring and dehazing performance of the network, we propose classification attention-based detection feature loss (CADF-Loss), which uses the image classification features to guide the network perform the detection feature enhancement more specifically. The experimental results in real navigation scenes show that D3-Net not only improves the detection accuracy of blurred and hazy images by 42%, but also has excellent detection efficiency, which is helpful to improve the navigation safety performance of smart ships.

The Construction and Evaluation of a Multi-Task Convolutional Neural Network for a Cone-Beam Computed-Tomography-Based Assessment of Implant Stability

Objectives: Assessing implant stability is integral to dental implant therapy. This study aimed to construct a multi-task cascade convolution neural network to evaluate implant stability using cone-beam computed tomography (CBCT). Methods: A dataset of 779 implant coronal section images was obtained from CBCT scans, and matching clinical information was used for the training and test datasets. We developed a multi-task cascade network based on CBCT to assess implant stability. We used the MobilenetV2-DeeplabV3+ semantic segmentation network, combined with an image processing algorithm in conjunction with prior knowledge, to generate the volume of interest (VOI) that was eventually used for the ResNet-50 classification of implant stability. The performance of the multitask cascade network was evaluated in a test set by comparing the implant stability quotient (ISQ), measured using an Osstell device. Results: The cascade network established in this study showed good prediction performance for implant stability classification. The binary, ternary, and quaternary ISQ classification test set accuracies were 96.13%, 95.33%, and 92.90%, with mean precisions of 96.20%, 95.33%, and 93.71%, respectively. In addition, this cascade network evaluated each implant's stability in only 3.76 s, indicating high efficiency. Conclusions: To our knowledge, this is the first study to present a CBCT-based deep learning approach CBCT to assess implant stability. The multi-task cascade network accomplishes a series of tasks related to implant denture segmentation, VOI extraction, and implant stability classification, and has good concordance with the ISQ.

In-situ monitoring system for weld geometry of laser welding based on multi-task convolutional neural network model

This paper presents a low-cost, robust, in-situ monitoring system for weld geometry that can achieve multi-task prediction. First, the system uses a low-cost CCD camera to monitor the melt pool in the laser keyhole welding process. Then, the proposed novel multi-task convolutional neural network (Multi-task CNN) model is used to simultaneously complete the two prediction tasks of weld depth and width. Furthermore, the learning process of the Multi-task CNN model is explored using a visual feature map approach and the robustness of the model is demonstrated. Compared with Support Vector Machine, K-Nearest Neighbor, Bayesian Ridge, Decision Tree, the proposed Multi-task CNN model has the highest prediction accuracy. The model predicts a mean absolute percentage error (MAPE, relative to ground truth) of 3.0% and 1.9% for weld depth and width. The in-situ monitoring results show that the system can achieve accurate predictions, and the average time-consuming of the system is 23.35 ms.

Improving Misfire Fault Diagnosis with Cascading Architectures via Acoustic Vehicle Characterization.

In a world dependent on road-based transportation, it is essential to understand automobiles. We propose an acoustic road vehicle characterization system as an integrated approach for using sound captured by mobile devices to enhance transparency and understanding of vehicles and their condition for non-expert users. We develop and implement novel deep learning cascading architectures, which we define as conditional, multi-level networks that process raw audio to extract highly granular insights for vehicle understanding. To showcase the viability of cascading architectures, we build a multi-task convolutional neural network that predicts and cascades vehicle attributes to enhance misfire fault detection. We train and test these models on a synthesized dataset reflecting more than 40 hours of augmented audio. Through cascading fuel type, engine configuration, cylinder count and aspiration type attributes, our cascading CNN achieves 87.0% test set accuracy on misfire fault detection which demonstrates margins of 8.0% and 1.7% over naïve and parallel CNN baselines. We explore experimental studies focused on acoustic features, data augmentation, and data reliability. Finally, we conclude with a discussion of broader implications, future directions, and application areas for this work.

Exploiting compressed sensing and polynomial-based progressive secret image sharing for visually secure image selection encryption with authentication

Image encryption is an effective method to protect private images by converting them into meaningless ones, but many current image encryption algorithms still have security and efficiency problems. This paper combines compressed sensing (CS) and secret image sharing (SIS) to propose a certifiable visually secure image selection encryption. Firstly, a selective encryption based on multi-task convolutional neural network (SE-MTCNN) is presented to distinguish and encrypt the sensitive and non-sensitive information of plain images. Subsequently, the SIS algorithm is used to decompose the resulting plain image into multiple shadow images to disperse the risk of cryptographic image loss and resist hacking attacks. Herein, a pseudo-convolutional sliding scrambling based on chaotic sequences driven by plaintext (PCSS-CSDP) is proposed to eliminate the correlation between adjacent image pixels. Finally, the IWT-M-embedding algorithm is used to embed shadow images into carrier images to get multiple visually meaningful cipher images. Additionally, a new authentication scheme based on bit-level cyclic shifts and dynamic combination (AS-BCSDC) is introduced, and the obtained authentication information is fused into cipher images to prevent the deception of illegal users and malicious tampering of cloud platform. Experiments demonstrate the effectiveness of our algorithm, and it may be applied for image secure communication.

Topology Optimization for Motor Using Multitask Convolutional Neural Network Under Multiple Current Conditions

This article proposes a new topology optimization (TO) method for a motor that leads to an optimized solution in less time under multiple current conditions while considering the torque ripple-order component. To consider the different characteristics, a multitask CNN is newly employed. The multitask CNN predicts multiple torque performances based on the current conditions and cross-sectional magnetic flux density distribution and applies it to the TO process. As a result, the average torque and harmonic components of the target motor improved by 8.9% and 48.2%, respectively, under multiple current conditions. Furthermore, the computational cost for TO was reduced by 95.1% using the proposed method, compared with that of conventional methods. Therefore, the proposed method enables fast optimization of torque ripple-order components under a wide range of current conditions.

Creating Semantic HD Maps From Aerial Imagery and Aggregated Vehicle Telemetry for Autonomous Vehicles

We present a new framework for creating lane-level detailed HD-maps at scale for autonomous vehicles (AVs). In order to overcome scaling challenges of ground survey-based HD-map creation, we propose a number of innovations by leveraging three data sources: high-resolution aerial imagery, aggregated vehicle telemetry, and a navigation map. We first divide map creation problem into several categories based on lane configurations. The road category is predicted in a supervised setting using aerial imagery, which is pre-processed by using aggregated vehicle telemetry without supervision. We utilized the navigation map to guide the process along the road network and used aerial imagery and aggregated vehicle telemetry to extract lane level features at each step. We propose a multi-task convolutional neural network (CNN) to predict road polygons, road-way centerline, number of lanes, lane and road edges using both the aerial imagery and the corresponding aggregated vehicle telemetry. The predicted road features for each image are then stitched along a road segment to construct the road and lane edge polylines, which are then used to predict lane marking and road edge types in a sliding window fashion along the road segment. The extracted features are finally utilized to calculate higher level features for each point in the HD-map. Our experimental results show that the proposed framework works well, offering a flexible solution for creating HD-maps for AVs at scale.

Feature-Level Attention-Guided Multitask CNN for Fault Diagnosis and Working Conditions Identification of Rolling Bearing.

Accurate and real-time fault diagnosis (FD) and working conditions identification (WCI) are the key to ensuring the safe operation of mechanical systems. We observe that there is a close correlation between the fault condition and the working condition in the vibration signal. Most of the intelligent FD methods only learn some features from the vibration signals and then use them to identify fault categories. They ignore the impact of working conditions on the bearing system, and such a single-task learning method cannot learn the complementary information contained in multiple related tasks. Therefore, this article is devoted to mining richer and complementary globally shared features from vibration signals to complete the FD and WCI of rolling bearings at the same time. To this end, we propose a novel multitask attention convolutional neural network (MTA-CNN) that can automatically give feature-level attention to specific tasks. The MTA-CNN consists of a global feature shared network (GFS-network) for learning globally shared features and K task-specific networks with feature-level attention module (FLA-module). This architecture allows the FLA-module to automatically learn the features of specific tasks from globally shared features, thereby sharing information among different tasks. We evaluated our method on the wheelset bearing data set and motor bearing data set. The results show that our method has a better performance than the state-of-the-art deep learning methods and strongly prove that our multitask learning mechanism can improve the results of each task.

Research on Classroom Emotion Recognition Algorithm Based on Visual Emotion Classification.

In this paper, we construct a classroom emotion recognition algorithm by classifying visual emotions for improving the quality of classroom teaching. We assign weights to the training images through an attention mechanism network and then add a designed loss function so that it can focus on the feature parts of face images that are not obscured and can characterize the target emotion, thus improving the accuracy of facial emotion recognition under obscuration. Analyze the salient expression features of classroom students and establish a classification criteria and criteria library. The videos of classroom students' facial expressions are collected, a multi-task convolutional neural network (MTCNN) is used for face detection and image segmentation, and the ones with better feature morphology are selected to build a standard database. A visual motion analysis method with the fusion of overall and local features of the image is proposed. To validate the effectiveness of the designed MTCNN model, two mainstream classification networks, VGG16 and ResNet18, were tested and compared with MTCNN by training on RAF-DB, masked dataset, and the classroom dataset constructed in this paper, and the final accuracy after training was 78.26% and 75.03% for ResNet18 and VGG16, respectively. The results show that the MTCNN proposed in this paper has a better recognition effect. The test results of the loss function also show that it can effectively improve the recognition accuracy, and the MTCNN model has an accuracy of 93.53% for recognizing students' facial emotions. Finally, the dataset is extended with the training method of expression features, and the experimental study shows that the method performs well and can carry out recognition effectively.

Dual-Channel Multi-Task CNN for No-Reference Screen Content Image Quality Assessment

Nowadays the problem of image quality assessment (IQA) for screen content images (SCIs) has become a research hotspot as they are ubiquitous in multimedia applications. Although the quality assessment of natural images (NIs) has been continuously developed in the past few decades, few NI-oriented IQA methods can be directly applied on SCIs due to different visual characteristics between them. In this paper, we present a no-reference quality prediction approach considering the content information of SCIs, which is based on dual-channel multi-task convolutional neural network. First, we segment a SCI into small patches and classify them as the textual patches and the pictorial patches. Then, we devise a novel dual-channel convolutional neural network (CNN) to predict the quality of textual patches and pictorial patches. Finally, we propose an effective adaptive weighting strategy for quality score aggregation. The proposed CNN is built on an end-to-end multi-task learning framework, which assists the SCI quality prediction task through the histogram of oriented gradient (HOG) feature prediction task to learn a better mapping between the input patch and its quality score. The adaptive weighting strategy further improves the representation ability of each SCI patch. Experimental results on two largest SCI-oriented databases demonstrate that the proposed method outperforms most of the state-of-the-art no-reference IQA methods and the full-reference IQA methods.