Online supervised attention-based recurrent depth estimation from monocular video.
Ontology highlight
ABSTRACT:
SUBMITTER: Maslov D
PROVIDER: S-EPMC7924529 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:This paper tackles the challenge of accurate depth estimation from monocular laparoscopic images in dynamic surgical environments. The lack of reliable ground truth due to inconsistencies within these images makes this a complex task. Further complicating the learning process is the presence of noise elements like bleeding and smoke. We propose a model learning framework that uses a generic laparoscopic surgery video dataset for training, aimed at achieving precise monocular depth estimation in dynamic surgical settings. The architecture employs binocular disparity confidence information as a self-supervisory signal, along with the disparity information from a stereo laparoscope. Our method ensures robust learning amidst outliers, influenced by tissue deformation, smoke, and surgical instruments, by utilizing a unique loss function. This function adjusts the selection and weighting of depth data for learning based on their given confidence. We trained the model using the Hamlyn Dataset and verified it with Hamlyn Dataset test data and a static dataset. The results show exceptional generalization performance and efficacy for various scene dynamics, laparoscope types, and surgical sites.
| S-EPMC10505201 | biostudies-literature
Project description:Many alternative approaches for 3D object detection using a singular camera have been studied instead of leveraging high-precision 3D LiDAR sensors incurring a prohibitive cost. Recently, we proposed a novel approach for 3D object detection by employing a ground plane model that utilizes geometric constraints named GAC3D to improve the results of the deep-based detector. GAC3D adopts an adaptive depth convolution to replace the traditional 2D convolution to deal with the divergent context of the image's feature, leading to a significant improvement in both training convergence and testing accuracy on the KITTI 3D object detection benchmark. This article presents an alternative architecture named eGAC3D that adopts a revised depth adaptive convolution with variant guidance to improve detection accuracy. Additionally, eGAC3D utilizes the pixel adaptive convolution to leverage the depth map to guide our model for detection heads instead of using an external depth estimator like other methods leading to a significant reduction of time inference. The experimental results on the KITTI benchmark show that our eGAC3D outperforms not only our previous GAC3D but also many existing monocular methods in terms of accuracy and inference time. Moreover, we deployed and optimized the proposed eGAC3D framework on an embedded platform with a low-cost GPU. To the best of the authors' knowledge, we are the first to develop a monocular 3D detection framework on embedded devices. The experimental results on Jetson Xavier NX demonstrate that our proposed method can achieve nearly real-time performance with appropriate accuracy even with the modest hardware resource.
| S-EPMC9680869 | biostudies-literature
Project description:Today, we human beings are facing with high-quality virtual world of a completely new nature. For example, we have a digital display consisting of a high enough resolution that we cannot distinguish from the real world. However, little is known how such high-quality representation contributes to the sense of realness, especially to depth perception. What is the neural mechanism of processing such fine but virtual representation? Here, we psychophysically and physiologically examined the relationship between stimulus resolution and depth perception, with using luminance-contrast (shading) as a monocular depth cue. As a result, we found that a higher resolution stimulus facilitates depth perception even when the stimulus resolution difference is undetectable. This finding is against the traditional cognitive hierarchy of visual information processing that visual input is processed continuously in a bottom-up cascade of cortical regions that analyze increasingly complex information such as depth information. In addition, functional magnetic resonance imaging (fMRI) results reveal that the human middle temporal (MT+) plays a significant role in monocular depth perception. These results might provide us with not only the new insight of our neural mechanism of depth perception but also the future progress of our neural system accompanied by state-of- the-art technologies.
| S-EPMC4202214 | biostudies-literature
Project description:Intercepting and avoiding moving objects requires accurate motion-in-depth (MID) perception. Such motion can be estimated based on both binocular and monocular cues. Because previous studies largely characterized sensitivity to these cues individually, their relative contributions to MID perception remain unclear. Here we measured sensitivity to binocular, monocular, and combined cue MID stimuli using a motion coherence paradigm. We first confirmed prior reports of substantial variability in binocular MID cue sensitivity across the visual field. The stimuli were matched for eccentricity and speed, suggesting that this variability has a neural basis. Second, we determined that monocular MID cue sensitivity also varied considerably across the visual field. A major component of this variability was geometric: An MID stimulus produces the largest motion signals in the eye contralateral to its visual field location. This resulted in better monocular discrimination performance when the contralateral rather than ipsilateral eye was stimulated. Third, we found that monocular cue sensitivity generally exceeded, and was independent of, binocular cue sensitivity. Finally, contralateral monocular cue sensitivity was found to be a strong predictor of combined cue sensitivity. These results reveal distinct factors constraining the contributions of binocular and monocular cues to three-dimensional motion perception.
| S-EPMC6402382 | biostudies-literature
Project description:Weakly supervised object localization tasks remain challenging to identify and segment an entire object rather than only discriminative parts of the object. To tackle this problem, corruption-based approaches have been devised, which involve the training of non-discriminative regions by corrupting (e.g., erasing) the input images or intermediate feature maps. However, this approach requires an additional hyperparameter, the corrupting threshold, to determine the degree of corruption and can unfavorably disrupt training. It also tends to localize object regions coarsely. In this paper, we propose a novel approach, Module of Axis-based Nexus Attention (MoANA), which helps to adaptively activate less discriminative regions along with the class-discriminative regions without an additional hyperparameter, and elaborately localizes an entire object. Specifically, MoANA consists of three mechanisms (1) triple-view attentions representation, (2) attentions expansion, and (3) features calibration mechanism. Unlike other attention-based methods that train a coarse attention map with the same values across elements in feature maps, MoANA trains fine-grained values in an attention map by assigning different attention values to each element. We validated MoANA by comparing it with various methods. We also analyzed the effect of each component in MoANA and visualized attention maps to provide insights into the calibration.
| S-EPMC10616293 | biostudies-literature
Project description:ObjectivesReliable determination of cochlear implant electrode positions shows promise for clinical applications, including anatomy-based fitting of audio processors or monitoring of electrode migration during follow-up. Currently, electrode positioning is measured using radiography. The primary objective of this study is to extend and validate an impedance-based method for estimating electrode insertion depths, which could serve as a radiation-free and cost-effective alternative to radiography. The secondary objective is to evaluate the reliability of the estimation method in the postoperative follow-up over several months.DesignThe ground truth insertion depths were measured from postoperative computed tomography scans obtained from the records of 56 cases with an identical lateral wall electrode array. For each of these cases, impedance telemetry records were retrieved starting from the day of implantation up to a maximum observation period of 60 mo. Based on these recordings, the linear and angular electrode insertion depths were estimated using a phenomenological model. The estimates obtained were compared with the ground truth values to calculate the accuracy of the model.ResultsAnalysis of the long-term recordings using a linear mixed-effects model showed that postoperative tissue resistances remained stable throughout the follow-up period, except for the two most basal electrodes, which increased significantly over time (electrode 11: ~10 Ω/year, electrode 12: ~30 Ω/year). Inferred phenomenological models from early and late impedance telemetry recordings were not different. The insertion depth of all electrodes was estimated with an absolute error of 0.9 mm ± 0.6 mm or 22° ± 18° angle (mean ± SD).ConclusionsInsertion depth estimations of the model were reliable over time when comparing two postoperative computed tomography scans of the same ear. Our results confirm that the impedance-based position estimation method can be applied to postoperative impedance telemetry recordings. Future work needs to address extracochlear electrode detection to increase the performance of the method.
| S-EPMC10583924 | biostudies-literature
Project description:Constructing ontology manually is a time-consuming, error-prone, and tedious task. We present SSCO, a self-supervised learning based chinese ontology, which contains about 255 thousand concepts, 5 million entities, and 40 million facts. We explore the three largest online Chinese encyclopedias for ontology learning and describe how to transfer the structured knowledge in encyclopedias, including article titles, category labels, redirection pages, taxonomy systems, and InfoBox modules, into ontological form. In order to avoid the errors in encyclopedias and enrich the learnt ontology, we also apply some machine learning based methods. First, we proof that the self-supervised machine learning method is practicable in Chinese relation extraction (at least for synonymy and hyponymy) statistically and experimentally and train some self-supervised models (SVMs and CRFs) for synonymy extraction, concept-subconcept relation extraction, and concept-instance relation extraction; the advantages of our methods are that all training examples are automatically generated from the structural information of encyclopedias and a few general heuristic rules. Finally, we evaluate SSCO in two aspects, scale and precision; manual evaluation results show that the ontology has excellent precision, and high coverage is concluded by comparing SSCO with other famous ontologies and knowledge bases; the experiment results also indicate that the self-supervised models obviously enrich SSCO.
| S-EPMC3970055 | biostudies-other
Project description:In natural contexts, sensory processing and motor output are closely coupled, which is reflected in the fact that many brain areas contain both sensory and movement signals. However, standard reductionist paradigms decouple sensory decisions from their natural motor consequences, and head-fixation prevents the natural sensory consequences of self-motion. In particular, movement through the environment provides a number of depth cues beyond stereo vision that are poorly understood. To study the integration of visual processing and motor output in a naturalistic task, we investigated distance estimation in freely moving mice. We found that mice use vision to accurately jump across a variable gap, thus directly coupling a visual computation to its corresponding ethological motor output. Monocular eyelid suture did not affect gap jumping success, thus mice can use cues that do not depend on binocular disparity and stereo vision. Under monocular conditions, mice altered their head positioning and performed more vertical head movements, consistent with a shift from using stereopsis to other monocular cues, such as motion or position parallax. Finally, optogenetic suppression of primary visual cortex impaired task performance under both binocular and monocular conditions when optical fiber placement was localized to binocular or monocular zone V1, respectively. Together, these results show that mice can use monocular cues, relying on visual cortex, to accurately judge distance. Furthermore, this behavioral paradigm provides a foundation for studying how neural circuits convert sensory information into ethological motor output.
| S-EPMC9489205 | biostudies-literature
Project description:Human gait analysis is often conducted in clinical and basic research, but many common approaches (e.g., three-dimensional motion capture, wearables) are expensive, immobile, data-limited, and require expertise. Recent advances in video-based pose estimation suggest potential for gait analysis using two-dimensional video collected from readily accessible devices (e.g., smartphones). To date, several studies have extracted features of human gait using markerless pose estimation. However, we currently lack evaluation of video-based approaches using a dataset of human gait for a wide range of gait parameters on a stride-by-stride basis and a workflow for performing gait analysis from video. Here, we compared spatiotemporal and sagittal kinematic gait parameters measured with OpenPose (open-source video-based human pose estimation) against simultaneously recorded three-dimensional motion capture from overground walking of healthy adults. When assessing all individual steps in the walking bouts, we observed mean absolute errors between motion capture and OpenPose of 0.02 s for temporal gait parameters (i.e., step time, stance time, swing time and double support time) and 0.049 m for step lengths. Accuracy improved when spatiotemporal gait parameters were calculated as individual participant mean values: mean absolute error was 0.01 s for temporal gait parameters and 0.018 m for step lengths. The greatest difference in gait speed between motion capture and OpenPose was less than 0.10 m s-1. Mean absolute error of sagittal plane hip, knee and ankle angles between motion capture and OpenPose were 4.0°, 5.6° and 7.4°. Our analysis workflow is freely available, involves minimal user input, and does not require prior gait analysis expertise. Finally, we offer suggestions and considerations for future applications of pose estimation for human gait analysis.
| S-EPMC8099131 | biostudies-literature
Project description:Video deblurring aims at removing the motion blur caused by the movement of objects or camera shake. Traditional video deblurring methods have mainly focused on frame-based deblurring, which takes only blurry frames as the input to produce sharp frames. However, frame-based deblurring has shown poor picture quality in challenging cases of video restoration where severely blurred frames are provided as the input. To overcome this issue, recent studies have begun to explore the event-based approach, which uses the event sequence captured by an event camera for motion deblurring. Event cameras have several advantages compared to conventional frame cameras. Among these advantages, event cameras have a low latency in imaging data acquisition (0.001 ms for event cameras vs. 10 ms for frame cameras). Hence, event data can be acquired at a high acquisition rate (up to one microsecond). This means that the event sequence contains more accurate motion information than video frames. Additionally, event data can be acquired with less motion blur. Due to these advantages, the use of event data is highly beneficial for achieving improvements in the quality of deblurred frames. Accordingly, the results of event-based video deblurring are superior to those of frame-based deblurring methods, even for severely blurred video frames. However, the direct use of event data can often generate visual artifacts in the final output frame (e.g., image noise and incorrect textures), because event data intrinsically contain insufficient textures and event noise. To tackle this issue in event-based deblurring, we propose a two-stage coarse-refinement network by adding a frame-based refinement stage that utilizes all the available frames with more abundant textures to further improve the picture quality of the first-stage coarse output. Specifically, a coarse intermediate frame is estimated by performing event-based video deblurring in the first-stage network. A residual hint attention (RHA) module is also proposed to extract useful attention information from the coarse output and all the available frames. This module connects the first and second stages and effectively guides the frame-based refinement of the coarse output. The final deblurred frame is then obtained by refining the coarse output using the residual hint attention and all the available frame information in the second-stage network. We validated the deblurring performance of the proposed network on the GoPro synthetic dataset (33 videos and 4702 frames) and the HQF real dataset (11 videos and 2212 frames). Compared to the state-of-the-art method (D2Net), we achieved a performance improvement of 1 dB in PSNR and 0.05 in SSIM on the GoPro dataset, and an improvement of 1.7 dB in PSNR and 0.03 in SSIM on the HQF dataset.
| S-EPMC10056412 | biostudies-literature