Comparison between two- and three-dimensional scoring of zebrafish response to psychoactive drugs: identifying when three-dimensional analysis is needed

PeerJ ◽

10.7717/peerj.7893 ◽

2019 ◽

Vol 7 ◽

pp. e7893 ◽

Cited By ~ 2

Author(s):

Simone Macrì ◽

Romain J.G. Clément ◽

Chiara Spinello ◽

Maurizio Porfiri

Keyword(s):

Time Course ◽

False Negative ◽

Three Dimensional ◽

Ground Truth ◽

Behavioral Pharmacology ◽

Experimental Session ◽

Behavioral Repertoire ◽

Ground Truth Data ◽

Negative Findings ◽

Data Yield

Zebrafish (Danio rerio) have recently emerged as a valuable laboratory species in the field of behavioral pharmacology, where they afford rapid and precise high-throughput drug screening. Although the behavioral repertoire of this species manifests along three-dimensional (3D), most of the efforts in behavioral pharmacology rely on two-dimensional (2D) projections acquired from a single overhead or front camera. We recently showed that, compared to a 3D scoring approach, 2D analyses could lead to inaccurate claims regarding individual and social behavior of drug-free experimental subjects. Here, we examined whether this conclusion extended to the field of behavioral pharmacology by phenotyping adult zebrafish, acutely exposed to citalopram (30, 50, and 100 mg/L) or ethanol (0.25%, 0.50%, and 1.00%), in the novel tank diving test over a 6-min experimental session. We observed that both compounds modulated the time course of general locomotion and anxiety-related profiles, the latter being represented by specific behaviors (erratic movements and freezing) and avoidance of anxiety-eliciting areas of the test tank (top half and distance from the side walls). We observed that 2D projections of 3D trajectories (ground truth data) may introduce a source of unwanted variation in zebrafish behavioral phenotyping. Predictably, both 2D views underestimate absolute levels of general locomotion. Additionally, while data obtained from a camera positioned on top of the experimental tank are similar to those obtained from a 3D reconstruction, 2D front view data yield false negative findings.

Enhancing unsupervised video-based vehicle tracking and modeling for traffic data collection

Canadian Journal of Civil Engineering ◽

10.1139/cjce-2019-0087 ◽

2020 ◽

Vol 47 (8) ◽

pp. 982-997

Author(s):

Mohamed H. Zaki ◽

Tarek Sayed ◽

Moataz Billeh

Keyword(s):

Data Collection ◽

Video Analysis ◽

Three Dimensional ◽

Ground Truth ◽

Traffic Data ◽

Ground Truth Data ◽

The Road ◽

Traffic Data Collection ◽

Bounding Boxes ◽

Automated Video Analysis

Video-based traffic analysis is a leading technology for streamlining transportation data collection. With traffic records from video cameras, unsupervised automated video analysis can detect various vehicle measures such as vehicle spatial coordinates and subsequently lane positions, speed, and other dynamic measures without the need of any physical interconnections to the road infrastructure. This paper contributes to the unsupervised automated video analysis by addressing two main shortcomings of the approach. The first objective is to alleviate tracking problems of over-segmentation and over-grouping by integrating region-based detection with feature-based tracking. This information, when combined with spatiotemporal constraints of grouping, can reduce the effects of these problems. This fusion approach offers a superior decision procedure for grouping objects and discriminating between trajectories of objects. The second objective is to model three-dimensional bounding boxes for the vehicles, leading to a better estimate of their geometry and consequently accurate measures of their position and travel information. This improvement leads to more precise measurement of traffic parameters such as average speed, gap time, and headway. The paper describes the various steps of the proposed improvements. It evaluates the effectiveness of the refinement process on data collected from traffic cameras in three different locations in Canada and validates the results with ground truth data. It illustrates the effectiveness of the improved unsupervised automated video analysis with a case study on 10 h of traffic data collection such as volume and headway measurements.

Transfer Change Rules from Recurrent Fully Convolutional Networks for Hyperspectral Unmanned Aerial Vehicle Images without Ground Truth Data

Remote Sensing ◽

10.3390/rs12071099 ◽

2020 ◽

Vol 12 (7) ◽

pp. 1099 ◽

Cited By ~ 2

Author(s):

Ahram Song ◽

Yongil Kim

Keyword(s):

Unmanned Aerial Vehicle ◽

Three Dimensional ◽

Ground Truth ◽

Training Data ◽

Fine Tuning ◽

Target Domain ◽

Ground Truth Data ◽

Convolutional Networks ◽

Fully Convolutional Networks ◽

Aerial Vehicle

Change detection (CD) networks based on supervised learning have been used in diverse CD tasks. However, such supervised CD networks require a large amount of data and only use information from current images. In addition, it is time consuming to manually acquire the ground truth data for newly obtained images. Here, we proposed a novel method for CD in case of a lack of training data in an area near by another one with the available ground truth data. The proposed method automatically entails generating training data and fine-tuning the CD network. To detect changes in target images without ground truth data, the difference images were generated using spectral similarity measure, and the training data were selected via fuzzy c-means clustering. Recurrent fully convolutional networks with multiscale three-dimensional filters were used to extract objects of various sizes from unmanned aerial vehicle (UAV) images. The CD network was pre-trained on labeled source domain data; then, the network was fine-tuned on target images using generated training data. Two further CD networks were trained with a combined weighted loss function. The training data in the target domain were iteratively updated using he prediction map of the CD network. Experiments on two hyperspectral UAV datasets confirmed that the proposed method is capable of transferring change rules and improving CD results based on training data extracted in an unsupervised way.

The Zurich urban micro aerial vehicle dataset

The International Journal of Robotics Research ◽

10.1177/0278364917702237 ◽

2017 ◽

Vol 36 (3) ◽

pp. 269-273 ◽

Cited By ~ 30

Author(s):

András L Majdik ◽

Charles Till ◽

Davide Scaramuzza

Keyword(s):

Three Dimensional ◽

Ground Truth ◽

Ground Level ◽

Urban Environments ◽

Sensor Data ◽

Measurement Unit ◽

Reconstruction Algorithms ◽

Ground Truth Data ◽

Micro Aerial Vehicle ◽

Aerial Vehicle

This paper presents a dataset recorded on-board a camera-equipped micro aerial vehicle flying within the urban streets of Zurich, Switzerland, at low altitudes (i.e. 5–15 m above the ground). The 2 km dataset consists of time synchronized aerial high-resolution images, global position system and inertial measurement unit sensor data, ground-level street view images, and ground truth data. The dataset is ideal to evaluate and benchmark appearance-based localization, monocular visual odometry, simultaneous localization and mapping, and online three-dimensional reconstruction algorithms for micro aerial vehicles in urban environments.

Low-Cost Three-Dimensional Modeling of Crop Plants

Sensors ◽

10.3390/s19132883 ◽

2019 ◽

Vol 19 (13) ◽

pp. 2883 ◽

Cited By ~ 8

Author(s):

Jorge Martinez-Guanter ◽

Ángela Ribeiro ◽

Gerassimos G. Peteinatos ◽

Manuel Pérez-Ruiz ◽

Roland Gerhards ◽

...

Keyword(s):

Low Cost ◽

Three Dimensional ◽

Ground Truth ◽

Reconstruction Algorithm ◽

3D Models ◽

Crop Plants ◽

Plant Phenotyping ◽

Area Index ◽

Ground Truth Data ◽

The Creation

Plant modeling can provide a more detailed overview regarding the basis of plant development throughout the life cycle. Three-dimensional processing algorithms are rapidly expanding in plant phenotyping programmes and in decision-making for agronomic management. Several methods have already been tested, but for practical implementations the trade-off between equipment cost, computational resources needed and the fidelity and accuracy in the reconstruction of the end-details needs to be assessed and quantified. This study examined the suitability of two low-cost systems for plant reconstruction. A low-cost Structure from Motion (SfM) technique was used to create 3D models for plant crop reconstruction. In the second method, an acquisition and reconstruction algorithm using an RGB-Depth Kinect v2 sensor was tested following a similar image acquisition procedure. The information was processed to create a dense point cloud, which allowed the creation of a 3D-polygon mesh representing every scanned plant. The selected crop plants corresponded to three different crops (maize, sugar beet and sunflower) that have structural and biological differences. The parameters measured from the model were validated with ground truth data of plant height, leaf area index and plant dry biomass using regression methods. The results showed strong consistency with good correlations between the calculated values in the models and the ground truth information. Although, the values obtained were always accurately estimated, differences between the methods and among the crops were found. The SfM method showed a slightly better result with regard to the reconstruction the end-details and the accuracy of the height estimation. Although the use of the processing algorithm is relatively fast, the use of RGB-D information is faster during the creation of the 3D models. Thus, both methods demonstrated robust results and provided great potential for use in both for indoor and outdoor scenarios. Consequently, these low-cost systems for 3D modeling are suitable for several situations where there is a need for model generation and also provide a favourable time-cost relationship.

Lung nodule detection in chest X-rays using synthetic ground-truth data comparing CNN-based diagnosis to human performance

Scientific Reports ◽

10.1038/s41598-021-94750-z ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Manuel Schultheiss ◽

Philipp Schmette ◽

Jannis Bodden ◽

Juliane Aichele ◽

Christina Müller-Leisse ◽

...

Keyword(s):

False Positive ◽

Human Performance ◽

False Negative ◽

Lung Nodule ◽

Synthetic Data ◽

Ground Truth ◽

Ct Scans ◽

X Rays ◽

Ground Truth Data ◽

Cad System

AbstractWe present a method to generate synthetic thorax radiographs with realistic nodules from CT scans, and a perfect ground truth knowledge. We evaluated the detection performance of nine radiologists and two convolutional neural networks in a reader study. Nodules were artificially inserted into the lung of a CT volume and synthetic radiographs were obtained by forward-projecting the volume. Hence, our framework allowed for a detailed evaluation of CAD systems’ and radiologists’ performance due to the availability of accurate ground-truth labels for nodules from synthetic data. Radiographs for network training (U-Net and RetinaNet) were generated from 855 CT scans of a public dataset. For the reader study, 201 radiographs were generated from 21 nodule-free CT scans with altering nodule positions, sizes and nodule counts of inserted nodules. Average true positive detections by nine radiologists were 248.8 nodules, 51.7 false positive predicted nodules and 121.2 false negative predicted nodules. The best performing CAD system achieved 268 true positives, 66 false positives and 102 false negatives. Corresponding weighted alternative free response operating characteristic figure-of-merits (wAFROC FOM) for the radiologists range from 0.54 to 0.87 compared to a value of 0.81 (CI 0.75–0.87) for the best performing CNN. The CNN did not perform significantly better against the combined average of the 9 readers (p = 0.49). Paramediastinal nodules accounted for most false positive and false negative detections by readers, which can be explained by the presence of more tissue in this area.

Assessment of Fringe Pattern Decomposition with a Cross-Correlation Index for Phase Retrieval in Fringe Projection 3D Measurements

Sensors ◽

10.3390/s18103578 ◽

2018 ◽

Vol 18 (10) ◽

pp. 3578 ◽

Cited By ~ 1

Author(s):

Xinjun Zhu ◽

Limei Song ◽

Hongyi Wang ◽

Qinghua Guo

Keyword(s):

Phase Retrieval ◽

Cross Correlation ◽

Three Dimensional ◽

Ground Truth ◽

Fringe Projection ◽

Model Parameters ◽

Correlation Index ◽

Transform Method ◽

Ground Truth Data ◽

Fringe Patterns

Phase retrieval from single frame projection fringe patterns, a fundamental and challenging problem in fringe projection measurement, attracts wide attention and various new methods have emerged to address this challenge. Many phase retrieval methods are based on the decomposition of fringe patterns into a background part and a fringe part, and then the phase is obtained from the decomposed fringe part. However, the decomposition results are subject to the selection of model parameters, which is usually performed manually by trial and error due to the lack of decomposition assessment rules under a no ground truth data situation. In this paper, we propose a cross-correlation index to assess the decomposition and phase retrieval results without the need of ground truth data. The feasibility of the proposed metric is verified by simulated and real fringe patterns with the well-known Fourier transform method and recently proposed Shearlet transform method. This work contributes to the automatic phase retrieval and three-dimensional (3D) measurement with less human intervention, and can be potentially employed in other fields such as phase retrieval in digital holography.

Towards Single 2D Image-Level Self-Supervision for 3D Human Pose and Shape Estimation

Applied Sciences ◽

10.3390/app11209724 ◽

2021 ◽

Vol 11 (20) ◽

pp. 9724

Author(s):

Junuk Cha ◽

Muhammad Saqlain ◽

Changhwa Lee ◽

Seongyeong Lee ◽

Seungeun Lee ◽

...

Keyword(s):

Large Scale ◽

Three Dimensional ◽

Ground Truth ◽

Small Subset ◽

Learning Approaches ◽

Shape Estimation ◽

Camera System ◽

Ground Truth Data ◽

Human Pose ◽

2D Images

Three-dimensional human pose and shape estimation is an important problem in the computer vision community, with numerous applications such as augmented reality, virtual reality, human computer interaction, and so on. However, training accurate 3D human pose and shape estimators based on deep learning approaches requires a large number of images and corresponding 3D ground-truth pose pairs, which are costly to collect. To relieve this constraint, various types of weakly or self-supervised pose estimation approaches have been proposed. Nevertheless, these methods still involve supervision signals, which require effort to collect, such as unpaired large-scale 3D ground truth data, a small subset of 3D labeled data, video priors, and so on. Often, they require installing equipment such as a calibrated multi-camera system to acquire strong multi-view priors. In this paper, we propose a self-supervised learning framework for 3D human pose and shape estimation that does not require other forms of supervision signals while using only single 2D images. Our framework inputs single 2D images, estimates human 3D meshes in the intermediate layers, and is trained to solve four types of self-supervision tasks (i.e., three image manipulation tasks and one neural rendering task) whose ground-truths are all based on the single 2D images themselves. Through experiments, we demonstrate the effectiveness of our approach on 3D human pose benchmark datasets (i.e., Human3.6M, 3DPW, and LSP), where we present the new state-of-the-art among weakly/self-supervised methods.

A deep learning based framework for the registration of three dimensional multi-modal medical images of the head

Scientific Reports ◽

10.1038/s41598-021-81044-7 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Kh Tohidul Islam ◽

Sudanthi Wijewickrema ◽

Stephen O’Leary

Keyword(s):

Deep Learning ◽

Ethical Issues ◽

Medical Images ◽

Three Dimensional ◽

Synthetic Data ◽

Ground Truth ◽

Ground Truth Data ◽

Moving Images ◽

Conventional Machine ◽

Manual Input

AbstractImage registration is a fundamental task in image analysis in which the transform that moves the coordinate system of one image to another is calculated. Registration of multi-modal medical images has important implications for clinical diagnosis, treatment planning, and image-guided surgery as it provides the means of bringing together complimentary information obtained from different image modalities. However, since different image modalities have different properties due to their different acquisition methods, it remains a challenging task to find a fast and accurate match between multi-modal images. Furthermore, due to reasons such as ethical issues and need for human expert intervention, it is difficult to collect a large database of labelled multi-modal medical images. In addition, manual input is required to determine the fixed and moving images as input to registration algorithms. In this paper, we address these issues and introduce a registration framework that (1) creates synthetic data to augment existing datasets, (2) generates ground truth data to be used in the training and testing of algorithms, (3) registers (using a combination of deep learning and conventional machine learning methods) multi-modal images in an accurate and fast manner, and (4) automatically classifies the image modality so that the process of registration can be fully automated. We validate the performance of the proposed framework on CT and MRI images of the head obtained from a publicly available registration database.

Modeling the Three-Dimensional Chromatin Structure from Hi-C Data with Transfer Learning

10.1101/2021.12.15.472387 ◽

2021 ◽

Author(s):

Tristan Meynier Georges ◽

Maria Anna Rapsomaniki

Keyword(s):

Deep Learning ◽

Transfer Learning ◽

Chromatin Structure ◽

Structure Prediction ◽

3D Structure ◽

Three Dimensional ◽

Ground Truth ◽

Ground Truth Data ◽

Contact Frequency ◽

3D Chromatin Structure

Recent studies have revealed the importance of three-dimensional (3D) chromatin structure in the regulation of vital biological processes. Contrary to protein folding, no experimental procedure that can directly determine ground-truth 3D chromatin coordinates exists. Instead, chromatin conformation is studied implicitly using high-throughput chromosome conformation capture (Hi-C) methods that quantify the frequency of all pairwise chromatin contacts. Computational methods that infer the 3D chromatin structure from Hi-C data are thus unsupervised, and limited by the assumption that contact frequency determines Euclidean distance. Inspired by recent developments in deep learning, in this work we explore the idea of transfer learning to address the crucial lack of ground-truth data for 3D chromatin structure inference. We present a novel method, Transfer learning Encoder for CHromatin 3D structure prediction (TECH-3D) that combines transfer learning with creative data generation procedures to reconstruct chromatin structure. Our work outperforms previous deep learning attempts for chromatin structure inference and exhibits similar results as state-of-the-art algorithms on many tests, without making any assumptions on the relationship between contact frequencies and Euclidean distances. Above all, TECH-3D presents a highly creative and novel approach, paving the way for future deep learning models.

Inter-observer variability between readers of CT images: all for one and one for all

Digital Diagnostics ◽

10.17816/dd60622 ◽

2021 ◽

Vol 2 (2) ◽

pp. 105-118

Author(s):

Nikolas S. Kulberg ◽

Roman V. Reshetnikov ◽

Vladimir P. Novik ◽

Alexey B. Elizarov ◽

Maxim A. Gusev ◽

...

Keyword(s):

Diagnostic Accuracy ◽

Ct Scan ◽

False Negative ◽

Lung Nodule ◽

Ground Truth ◽

Observer Variability ◽

True Negative ◽

Inter Observer Variability ◽

Negative Findings ◽

Accuracy Increase

BACKGROUND: The markup of medical image datasets is based on the subjective interpretation of the observed entities by radiologists. There is currently no widely accepted protocol for determining ground truth based on radiologists reports. AIM: To assess the accuracy of radiologist interpretations and their agreement for the publicly available dataset CTLungCa-500, as well as the relationship between these parameters and the number of independent readers of CT scans. MATERIALS AND METHODS: Thirty-four radiologists took part in the dataset markup. The dataset included 536 patients who were at high risk of developing lung cancer. For each scan, six radiologists worked independently to create a report. After that, an arbitrator reviewed the lesions discovered by them. The number of true-positive, false-positive, true-negative, and false-negative findings was calculated for each reader to assess diagnostic accuracy. Further, the inter-observer variability was analyzed using the percentage agreement metric. RESULTS: An increase in the number of independent readers providing CT scan interpretations leads to accuracy increase associated with a decrease in agreement. The majority of disagreements were associated with the presence of a lung nodule in a specific site of the CT scan. CONCLUSION: If arbitration is provided, an increase in the number of independent initial readers can improve their combined accuracy. The experience and diagnostic accuracy of individual readers have no bearing on the quality of a crowd-tagging annotation. At four independent readings per CT scan, the optimal balance of markup accuracy and cost was achieved.