Real-time Detector Research Articles

FL-YOLOv8: Lightweight Object Detector Based on Feature Fusion

In recent years, anchor-free object detectors have become predominant in deep learning, the YOLOv8 model as a real-time object detector based on anchor-free frames is universal and influential, it efficiently detects objects across multiple scales. However, the generalization performance of the model is lacking, and the feature fusion within the neck module overly relies on its structural design and dataset size, and it is particularly difficult to localize and detect small objects. To address these issues, we propose the FL-YOLOv8 object detector, which is improved based on YOLOv8s. Firstly, we introduce the FSDI module in the neck, enhancing semantic information across all layers and incorporating rich detailed features through straightforward layer-hopping connections. This module integrates both high-level and low-level information to enhance the accuracy and efficiency of image detection. Meanwhile, the structure of the model was optimized and designed, and the LSCD module is constructed in the detection head; adopting a lightweight shared convolutional detection head reduces the number of parameters and computation of the model by 19% and 10%, respectively. Our model achieves a comprehensive performance of 45.5% on the COCO generalized dataset, surpassing the benchmark by 0.8 percentage points. To further validate the effectiveness of the method, experiments were also performed on specific domain urine sediment data (FCUS22), and the results on category detection also better justify the FL-YOLOv8 object detection algorithm.

Development of a Real-Time Wearable Humming Detector Device.

This study focuses on the development of a wearable real-time Humming Detector Device (HDD) aimed at enhancing the control of assistive devices through humming. As the need for portable user-friendly tools in assistive technology grows, the HDD offers a non-invasive solution to detect vocal cord vibrations. Vibrations, detected thanks to an accelerometer worn on the neck, are processed in real time using a Fast Fourier Transform (FFT) to identify specific humming frequencies, which are then translated into commands for controlling assistive devices via Bluetooth Low Energy (BLE) transmission. The device was tested with 13 healthy subjects to validate its potential and determine the optimal number of distinct commands that users can achieve. The HDD's portability and precision make it a promising alternative to traditional voice recognition systems, particularly for individuals with speech impairments.

Review of real time 2D dosimetry in external radiotherapy: Advancements and techniques

The objective of this paper is to provide a comprehensive review of the advancements and techniques in real time two-dimensional (2D) dosimetry for external radiation therapy with emphasis in vivo dosimetry and patient specific quality assurance. External radiation therapy plays a crucial role in cancer treatment, delivering high-energy radiation beams to target tumours while minimizing damage to surrounding healthy tissues. Accurate dosimetry, as both the measurement of the dose and its delivered location, is paramount to ensure effective treatment outcomes and minimize potential side effects.The planned content of this paper encompasses a thorough examination of the advancements made in 2D dosimetry techniques, including solid state and electronic systems. The evolution from traditional passive dosimetry to modern real time detectors, such as portal imaging, has revolutionized the field, offering enhanced precision, efficiency, and convenience. This review will discuss the principles, advantages, and limitations of these systems, along with their practical implementation and calibration procedures.Furthermore, the paper will highlight novel technologies, such as luminescence coatings, for quality assurance (QA) and real-time dose verification during treatment. The use of innovative materials and designs in dosemeters, including high spatial resolution detectors and tissue-equivalent phantoms, will also be explored. Additionally, the incorporation of advanced data analysis techniques, such as machine/deep learning algorithms, for dose reconstruction and QA will be addressed.The review will also explore the application of real time 2D dosimetry in modern clinical and pre-clinical modalities, including intensity-modulated radiation therapy and volumetric modulated arc therapy, stereotactic radiosurgery, image-guided radiation therapy, particle therapy, adaptive radiotherapy, electron and proton ultra-high dose rate therapy and very high energy electrons.By providing an up-to-date overview of the state-of-the-art in real time 2D dosimetry in vivo dosimetry and patient specific quality assurance, this paper aims to inform and guide professionals in the field, facilitating the adoption of cutting-edge techniques and improving the accuracy and safety of external radiotherapy treatments.

Sign language recognition using modified deep learning network and hybrid optimization: a hybrid optimizer (HO) based optimized CNNSa-LSTM approach.

Speech impairment limits a person's capacity for oral and auditory communication. Improvements in communication between the deaf and the general public can be progressed by a real-time sign language detector. Recent studies have contributed to make progress in motion and gesture identification processes using Deep Learning (DL) methods and computer vision. But the development of static and dynamic sign language recognition (SLR) models is still a challenging area of research. The difficulty is in obtaining an appropriate model that addresses the challenges of continuous signs that are independent of the signer. Different signers' speeds, durations, and many other factors make it challenging to create a model with high accuracy and continuity. This study mainly focused on SLR using a modified DL and hybrid optimization approach. Notably, spatial and geometric-based features are extracted via the Visual Geometry Group 16 (VGG16), and motion features are extracted using the optical flow approach. A new DL model, CNNSa-LSTM, is a combination of a Convolutional Neural Network (CNN), Self-Attention (SA), and Long-Short-Term Memory (LSTM) to identify sign language. This model is developed for feature extraction by combining CNNs for spatial analysis with SA mechanisms for focusing on relevant features, while LSTM effectively models temporal dependencies. The proposed CNNSa-LSTM model enhances performance in tasks involving complex, sequential data, such as sign language processing. Besides, a Hybrid Optimizer (HO) is proposed using the Hippopotamus Optimization Algorithm (HOA) and the Pathfinder Algorithm (PFA). The proposed model has been implemented in Python, and it has been evaluated over the existing models in terms of accuracy (98.7%), sensitivity (98.2%), precision (98.5%), Word Error Rate (WER) (0.131), Sign Error Rate (SER) (0.114), and Normalized Discounted Cumulative Gain (NDCG) (98%) as well. The proposed model has recorded the highest accuracy of 98.7%.

WGS-YOLO: A real-time object detector based on YOLO framework for autonomous driving

The safety and reliability of autonomous driving depends on the precision and efficiency of object detection systems. In this paper, a refined adaptation of the YOLO architecture (WGS-YOLO) is developed to improve the detection of pedestrians and vehicles. Specifically, its information fusion is enhanced by incorporating the Weighted Efficient Layer Aggregation Network (W-ELAN) module, an innovative dynamic weighted feature fusion module using channel shuffling. Meanwhile, the computational demands and parameters of the proposed WGS-YOLO are significantly reduced by employing the Space-to-Depth Convolution (SPD-Conv) and the Grouped Spatial Pyramid Pooling (GSPP) modules that have been strategically designed. The performance of our model is evaluated with the BDD100k and DAIR-V2X-V datasets. In terms of mean Average Precision (mAP0.5), the proposed model outperforms the baseline Yolov7 by 12%. Furthermore, extensive experiments are conducted to verify our analysis and the model’s robustness across diverse scenarios.

PPDM++: Parallel Point Detection and Matching for Fast and Accurate HOI Detection.

Human-Object Interaction (HOI) detection aims to understand human activities by detecting interaction triplets. Previous HOI detection methods adopt a two-stage instance-driven paradigm. Unfortunately, many non-interactive human-object pairs generated by the first stage are the main obstacle impeding HOI detectors from high efficiency and promising performance. To remedy this, we propose a novel top-down interaction-driven paradigm, detecting interactions first and bridging interactive human-object pairs through interactions. We formulate HOI as a point triplet human point, interaction point, object point and design a Parallel Point Detection and Matching (PPDM) framework. We further take advantage of two-stage methods and propose a novel framework, PPDM++, that detects the interactive human-object pairs by PPDM, then extracts region features for each pair to predict actions. The core of PPDM/PPDM++ is to convert the instance-driven bottom-up paradigm to an interaction-driven top-down paradigm, thus avoiding additional computation costs from traversing a tremendous number of non-interactive pairs. Benefiting from the advanced paradigm, PPDM/PPDM++ has achieved significant performance gains with high efficiency. PPDM-DLA-34 has achieved 19.94 mAP with 42 FPS as the first real-time HOI detector, and PPDM++-SwinB achieves 30.1 mAP with 17 FPS on HICO-DET dataset. We also built an application-oriented database named HOI-A, a supplement to the existing datasets.

Fluorescence of Intrinsic Milk Chromophores as a Novel Verification Method of UV-C Treatment of Milk.

The European Food Safety Authority (EFSA) has approved the use of a 1045 J/L UV-C dose as an adjunct to pasteurization to increase the shelf life and vitamin D3 content of milk. However, there are no verification methods analogous to the alkaline phosphatase test for pasteurized milk to ensure that the desired UV-C dose has been correctly applied. The aim is to develop a real-time in-line detector based on fluorescence spectroscopy. In this study, 22 different UV-C doses (ranging from 0 to 2000 J/L) were applied to milk to assess the impact of photooxidation on intrinsic photosensitive chromophores. Fluorescence spectroscopy (90°-angle) was employed as the method of analysis for monitoring the changes in the fluorescence spectra of chromophores in milk without sample pretreatment. Three important chromophore areas (CAs) were identified: CA 1 (riboflavin), CA 3 (vitamin A and dityrosine) and CA 4 (tryptophan), with statistically significant changes at around 1045 J/L and 1500 J/L. The findings of our preliminary study support our hypothesis that the fluorescence of intrinsic chromophores can be used as verification of the applied UV-C dose.

Quad Gaussian Networks for Vehicle Detection in Aerial Images.

Vehicle detection in remote sensing images is a crucial aspect of intelligent transportation systems. It plays an essential role in road planning, congestion control, and road construction in cities. However, detecting vehicles in remote sensing images is challenging due to their small size, high density, and noise. Most current detectors that perform well in conventional scenes fail to achieve better results in this context. Thus, we propose a quad-layer decoupled network to improve the algorithm's performance in detecting vehicles in remote sensing scenes. This is achieved by introducing modules such as a Group Focus downsampling structure, a quad-layer decoupled detector, and the GTAA label assignment method. Experiments demonstrate that the designed algorithm achieves a mean average precision (mAP) of 49.4 and operates at a speed of 3.0 ms on the RTX3090 within a multi-class vehicle detection dataset constructed based on the xView dataset. It outperforms various real-time detectors in terms of detection accuracy and speed.

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.

SW-YOLOX: A YOLOX-based real-time pedestrian detector with shift window-mixed attention mechanism

Pedestrian detection is a critical research area in computer vision with practical applications. This paper addresses this key topic by providing a novel lightweight model named Shift Window-YOLOX (SW-YOLOX). The purpose of SW-YOLOX is to significantly enhance the robustness and real-time performance of pedestrian detection under practical application requirements. The proposed method incorporates a novel Shift Window-Mixed Attention Mechanism (SW-MAM), which combines spatial and channel attention for effective feature extraction. In addition, we introduce a novel up-sampling layer, PatchExpandingv2, to enhance spatial feature representation while maintaining computational efficiency. Furthermore, we propose a novel Shift Window-Path Aggregation Feature Pyramid Network (SW-PAFPN) to integrate with the YOLOX detector, further enhancing feature extraction and the robustness of pedestrian detection. Experimental results validated on challenging datasets such as CrowdHuman, MOT17Det, and MOT20Det demonstrate the competitive performance of the proposed SW-YOLOX compared to state-of-the-art methods and its pedestrian detection performance in crowded and complex scenes.

DBHF: A double fusion backbone and bidirectional feature hybrid fusion detector for high-precision inspection of avionics solder joint defects

The quality of avionics thermistor wire solder joints has a critical impact on the performance of in-orbit spacecraft. However, it is still hard to simultaneously realize a high-precision inspection of the multi-scale internal and external defects of avionics solder joints. Herein, in combination with active infrared thermography technique, an innovative DBHF detector with a double fusion backbone and a learnable weighted multi-scale bidirectional feature hybrid fusion pyramid network (BdFPN) is demonstrated based on RT-DETR framework. Specifically, the double fusion backbone framework comprises with two lightweight convolutional neural network-based backbones. Compared with conventional single backbone, the double-backbone design forms a multi-level Resnet-like structure, helping to enrich the extracted multi-dimension feature information and retain rich detail information of small-scale defects. The complex learnable weighted multi-scale BdFPN is designed for cross fusion of deep features. One or more additional multi-scale inputs are added to each node involved in feature fusion to aggregate more features from multiple scales and strengthen the interaction between low-level details and high-level semantic information, enabling the detector to make full use of multi-level feature information. Meanwhile, a learnable weight coefficient is set on each node involved in feature fusion and the importance of the input to each node is estimated by training, so that the effective information can be deeply aggregated at different levels. Our DBHF detector achieves a 92.7 % mAP@0.5 on a self-made avionics solder joint dataset with a body size of only 27.2 MB and FPS of 149, outperforming all the typical advanced real-time detectors.

Object Detection in Multispectral Remote Sensing Images Based on Cross-Modal Cross-Attention.

In complex environments a single visible image is not good enough to perceive the environment, this paper proposes a novel dual-stream real-time detector designed for target detection in extreme environments such as nighttime and fog, which is able to efficiently utilise both visible and infrared images to achieve Fast All-Weatherenvironment sensing (FAWDet). Firstly, in order to allow the network to process information from different modalities simultaneously, this paper expands the state-of-the-art end-to-end detector YOLOv8, the backbone is expanded in parallel as a dual stream. Then, for purpose of avoid information loss in the process of network deepening, a cross-modal feature enhancement module is designed in this study, which enhances each modal feature by cross-modal attention mechanisms, thus effectively avoiding information loss and improving the detection capability of small targets. In addition, for the significant differences between modal features, this paper proposes a three-stage fusion strategy to optimise the feature integration through the fusion of spatial, channel and overall dimensions. It is worth mentioning that the cross-modal feature fusion module adopts an end-to-end training approach. Extensive experiments on two datasets validate that the proposed method achieves state-of-the-art performance in detecting small targets. The cross-modal real-time detector in this study not only demonstrates excellent stability and robust detection performance, but also provides a new solution for target detection techniques in extreme environments.

YOLO-FaceV2: A scale and occlusion aware face detector

In recent years, face detection algorithms based on deep learning have made great progress. Nevertheless, the effective utilization of face detectors for small and occlusion faces remains challenging, primarily stemming from the limitations in pixel information and the presence of missing features. In this paper, we propose a novel real-time face detector, YOLO-FaceV2, built upon the YOLOv5 architecture. Our approach introduces a Receptive Field Enhancement (RFE) module designed to extract multi-scale pixel information and augment the receptive field for accurately detecting small faces. To address issues related to face occlusion, we introduce an attention mechanism termed the Separated and Enhancement Attention Module (SEAM), which effectively focuses on the regions affected by occlusion. Furthermore, we propose a Slide Weight Function (SWF) to mitigate the imbalance between easy and hard samples. The experiments demonstrate that our YOLO-FaceV2 achieves performance exceeding the state-of-the-art on the WiderFace validation dataset. Source code and pre-trained model are available at https://github.com/Krasjet-Yu/YOLO-FaceV2.

EUAVDet: An Efficient and Lightweight Object Detector for UAV Aerial Images with an Edge-Based Computing Platform

Crafting an edge-based real-time object detector for unmanned aerial vehicle (UAV) aerial images is challenging because of the limited computational resources and the small size of detected objects. Existing lightweight object detectors often prioritize speed over detecting extremely small targets. To better balance this trade-off, this paper proposes an efficient and low-complexity object detector for edge computing platforms deployed on UAVs, termed EUAVDet (Edge-based UAV Object Detector). Specifically, an efficient feature downsampling module and a novel multi-kernel aggregation block are first introduced into the backbone network to retain more feature details and capture richer spatial information. Subsequently, an improved feature pyramid network with a faster ghost module is incorporated into the neck network to fuse multi-scale features with fewer parameters. Experimental evaluations on the VisDrone, SeaDronesSeeV2, and UAVDT datasets demonstrate the effectiveness and plug-and-play capability of our proposed modules. Compared with the state-of-the-art YOLOv8 detector, the proposed EUAVDet achieves better performance in nearly all the metrics, including parameters, FLOPs, mAP, and FPS. The smallest version of EUAVDet (EUAVDet-n) contains only 1.34 M parameters and achieves over 20 fps on the Jetson Nano. Our algorithm strikes a better balance between detection accuracy and inference speed, making it suitable for edge-based UAV applications.

X-CDNet: A real-time crosswalk detector based on YOLOX

As urban traffic safety becomes increasingly important, real-time crosswalk detection is playing a critical role in the transportation field. However, existing crosswalk detection algorithms must be improved in terms of accuracy and speed. This study proposes a real-time crosswalk detector called X-CDNet based on YOLOX. Based on the ConvNeXt basic module, we designed a new basic module called Reparameterizable Sparse Large-Kernel (RepSLK) convolution that can be used to expand the model’s receptive field without the addition of extra inference time. In addition, we created a new crosswalk dataset called CD9K, which is based on realistic driving scenes augmented by techniques such as synthetic rain and fog. The experimental results demonstrate that X-CDNet outperforms YOLOX in terms of both detection accuracy and speed. X-CDNet achieves a 93.3 AP50 and a real-time detection speed of 123 FPS.

A comparative study of a standard slit-to-agar sampler and a real-time bacterial detector

A comparison is presented of data from simultaneous measurements of airborne particles by a standard slit-to-agar sampler (STA) and of fluorescent particles by a real-time bacterial detector, Bio Aerosol Monitoring System (BAMS) during an evaluation regarding emitted airborne particles of a surgical clothing system in a test chamber. Data from simultaneous orientating measurements of airborne particles by a standard discrete particle counter are also discussed. The results show that in an environment with a low level of airborne particles including CFUs and man the only contamination source, there is a moderate correlation between the number of aerobic CFU and BAMS viable, i.e., fluorescence, particles, when the slit-to-agar sampler is registering/collecting within its detection level. Different environments need simultaneous measurements by an STA in aerobic CFU/m³ and by a BAMS in viable particles/m³ to set up the correlation between the two methods. Keywords: Airborne particles, controlled environment, CFU, fluorescent particles, real-time measurements, viable particles.

Easy Rocap: A Low-Cost and Easy-to-Use Motion Capture System for Drones

Fast and accurate pose estimation is essential for the local motion control of robots such as drones. At present, camera-based motion capture (Mocap) systems are mostly used by robots. However, this kind of Mocap system is easily affected by light noise and camera occlusion, and the cost of common commercial Mocap systems is high. To address these challenges, we propose Easy Rocap, a low-cost, open-source robot motion capture system, which can quickly and robustly capture the accurate position and orientation of the robot. Firstly, based on training a real-time object detector, an object-filtering algorithm using class and confidence is designed to eliminate false detections. Secondly, multiple-object tracking (MOT) is applied to maintain the continuity of the trajectories, and the epipolar constraint is applied to multi-view correspondences. Finally, the calibrated multi-view cameras are used to calculate the 3D coordinates of the markers and effectively estimate the 3D pose of the target robot. Our system takes in real-time multi-camera data streams, making it easy to integrate into the robot system. In the simulation scenario experiment, the average position estimation error of the method is less than 0.008 m, and the average orientation error is less than 0.65 degrees. In the real scenario experiment, we compared the localization results of our method with the advanced LiDAR-Inertial Simultaneous Localization and Mapping (SLAM) algorithm. According to the experimental results, SLAM generates drifts during turns, while our method can overcome the drifts and accumulated errors of SLAM, making the trajectory more stable and accurate. In addition, the pose estimation speed of our system can reach 30 Hz.

Toward Generating a New Cloud-Based Distributed Denial of Service (DDoS) Dataset and Cloud Intrusion Traffic Characterization

The distributed denial of service attack poses a significant threat to network security. Despite the availability of various methods for detecting DDoS attacks, the challenge remains in creating real-time detectors with minimal computational overhead. Additionally, the effectiveness of new detection methods depends heavily on well-constructed datasets. This paper addresses the critical DDoS dataset creation and evaluation domain, focusing on the cloud network. After conducting an in-depth analysis of 16 publicly available datasets, this research identifies 15 shortcomings across various dimensions, emphasizing the need for a new approach to dataset creation. Building upon this understanding, this paper introduces a new public DDoS dataset named BCCC-cPacket-Cloud-DDoS-2024. This dataset is meticulously crafted, addressing challenges identified in previous datasets through a cloud infrastructure featuring over eight benign user activities and 17 DDoS attack scenarios. Also, a Benign User Profiler (BUP) tool has been designed and developed to generate benign user network traffic based on a normal user behavior profile. We manually label the dataset and extract over 300 features from the network and transport layers of the traffic flows using NTLFlowLyzer. The experimental phase involves identifying an optimal feature set using three distinct algorithms: ANOVA, information gain, and extra tree. Finally, this paper proposes a multi-layered DDoS detection model and evaluates its performance using the generated dataset to cover the main issues of the traditional approaches.

Slim-neck by GSConv: a lightweight-design for real-time detector architectures

Slim-neck by GSConv: a lightweight-design for real-time detector architectures

Characterization of a flexible a-Si:H detector for in vivo dosimetry in therapeutic x-ray beams.

The increasing use of complex and high dose-rate treatments in radiation therapy necessitates advanced detectors to provide accurate dosimetry. Rather than relying on pre-treatment quality assurance (QA) measurements alone, many countries are now mandating the use of in vivo dosimetry, whereby a dosimeter is placed on the surface of the patient during treatment. Ideally, in vivo detectors should be flexible to conform to a patient's irregular surfaces. This study aims to characterize a novel hydrogenated amorphous silicon (a-Si:H) radiation detector for the dosimetry of therapeutic x-ray beams. The detectors are flexible as they are fabricated directly on a flexible polyimide (Kapton) substrate. The potential of this technology for application as a real-time flexible detector is investigated through a combined dosimetric and flexibility study. Measurements of fundamental dosimetric quantities were obtained including output factor (OF), dose rate dependence (DPP), energy dependence, percentage depth dose (PDD), and angular dependence. The response of the a-Si:H detectors investigated in this study are benchmarked directly against commercially available ionization chambers and solid-state diodes currently employed for QA practices. The a-Si:H detectors exhibit remarkable dose linearities in the direct detection of kV and MV therapeutic x-rays, with calibrated sensitivities ranging from (0.580±0.002) pC/cGy to (19.36±0.10) pC/cGy as a function of detector thickness, area, and applied bias. Regarding dosimetry, the a-Si:H detectors accurately obtained OF measurements that parallel commercially available detector solutions. The PDD response closely matched the expected profile as predicted via Geant4 simulations, a PTW Farmer ionization chamber and a PTW ROOS chamber. The most significant variation in the PDD performance was 5.67%, observed at a depth of 3mm for detectors operated unbiased. With an external bias, the discrepancy in PDD response from reference data was confined to±2.92% for all depths (surface to 250mm) in water-equivalent plastic. Very little angular dependence is displayed between irradiations at angles of 0° and 180°, with the most significant variation being a 7.71% decrease in collected charge at a 110° relative angle of incidence. Energy dependence and dose per pulse dependence are also reported, with results in agreement with the literature. Most notably, the flexibility of a-Si:H detectors was quantified for sample bending up to a radius of curvature of 7.98mm, where the recorded photosensitivity degraded by (-4.9±0.6)% of the initial device response when flat. It is essential to mention that this small bending radius is unlikely during in vivo patient dosimetry. In a more realistic scenario, with a bending radius of 15-20mm, the variation in detector response remained within±4%. After substantial bending, the detector's photosensitivity when returned to a flat condition was (99.1±0.5)% of the original response. This work successfully characterizes a flexible detector based on thin-film a-Si:H deposited on a Kapton substrate for applications in therapeutic x-ray dosimetry. The detectors exhibit dosimetric performances that parallel commercially available dosimeters, while also demonstrating excellent flexibility results.