Automatic Choroid Vascularity Index Calculation in Optical Coherence Tomography Images Low Contrast Sclerochoroidal Junction Using Deep Learning

Choroidal vascularity index (CVI) is a new biomarker defined for retinal optical coherence tomography (OCT) images for measuring and evaluating the choroidal vascular structure. CVI is the ratio of the choroidal luminal area (LA) to the total choroidal area (TCA). The automatic calculation of this index is important for ophthalmologists but has not yet been explored. In this study, we proposed a fully automated method based on deep learning for calculating CVI in three main steps: 1- segmentation of the choroidal boundary, 2- detection of the choroidal luminal vessels, and 3- computation of the CVI. The proposed method is evaluated in complex situations like the presence of diabetic retinopathy and pachychoroid spectrum. In pachychoroid spectrum, the choroid is thickened, and the boundary between choroid and sclera (sclerochoroidal junction) is blurred, which makes the segmentation more challenging. The proposed method is designed based on the U-Net model, and a new loss function is proposed to overcome the segmentation problems. The vascular LA is then calculated using Niblack’s local thresholding method, and the CVI value is finally computed. The experimental results for the segmentation stage with the best-performing model and the proposed loss function were used showed dice coefficients of 0.941 and 0.936 in diabetic retinopathy and pachychoroid spectrum patients, respectively. The unsigned boundary localization errors in the presence of diabetic retinopathy were 0.0020 and 0.0138 pixels for the BM boundary and sclerochoroidal junction, respectively. Similarly, the unsigned errors in the presence of pachychoroid spectrum were 0.0072 and 0.0254 pixels for BM and sclerochoroidal junction. The performance of the proposed method for calculating CVI was evaluated; the Bland-Altman plot indicated acceptable agreement between the values allocated by experts and the proposed method in the presence of diabetic retinopathy and pachychoroid spectrum.

Longitude and Latitude Based Received Signal Strength Difference Localization Models

In order to use the latitude and longitude coordinates for received signal strength difference (RSSD) localization, the errors of several spherical distance calculation methods and the error of arc length relative to string length were compared. The distance-calculation RSSD localization equations were established, including spherical accurate calculation RSSD, spherical approximate calculation RSSD, and normal cylindrical projection RSSD. And then, the optimization RSSD localization models based on geodetic coordinates and corresponding to the above equations were established, and the models were verified using the point by point search method with good convergence. The numerical results show there are a lot of weak localization areas for the RSSD localization networks lack of central stations with 4,5,6 stations. Among networks with central stations, there are only a small number of weak-localization areas for the concave 4 stations network, while there are no weak-localization areas for the networks composed of more stations. When the measurement errors and the additional losses of radio wave propagation are not considered, the localization errors of the spherical accurate model, the spherical approximate model and the equianglular projection model are very small, among which the second model has the shortest localization time. The localization errors of equidistance projection model and equal-area projection model are large, neither of which is suitable for the middle latitude and high latitude areas.

Enhancing Feature Point-Based Video Watermarking against Geometric Attacks with Template

As the Internet and communication technologies have developed quickly, the spread and usage of online video content have become easier, which results in major infringement problems. While video watermarking may be a viable solution for digital video content copyright protection, overcoming geometric attacks is a significant challenge. Although feature point-based watermarking algorithms are expected to be very resistant to these attacks, they are sensitive to feature region localization errors, resulting in poor watermark extraction accuracy. To solve this issue, we introduce the template to enhance the location accuracy of feature point-based watermarking. Furthermore, a scene change-based frame allocation method is presented, which arranges the template and the watermark to be embedded into different frames and eliminates their mutual interference, enhancing the performance of the proposed algorithm. According to the experimental results, our algorithm outperforms state-of-the-art methods in terms of robustness against geometric attacks under close imperceptibility.

Unreferenced spatial localization under monocular and dichoptic viewing conditions

Knowledge of eye position in the brain is critical for localization of objects in space. To investigate the accuracy and precision of eye position feedback in an unreferenced environment, subjects with normal ocular alignment attempted to localize briefly presented targets during monocular and dichoptic viewing. In the task, subjects’ used a computer mouse to position a response disk at the remembered location of the target. Under dichoptic viewing (with red (right eye)–green (left eye) glasses), target and response disks were presented to the same or alternate eyes, leading to four conditions [green target–green response cue (LL), green–red (LR), red–green (RL), and red–red (RR)]. Time interval between target and response disks was varied and localization errors were the difference between the estimated and real positions of the target disk. Overall, the precision of spatial localization (variance across trials) became progressively worse with time. Under dichoptic viewing, localization errors were significantly greater for alternate-eye trials as compared to same-eye trials and were correlated to the average phoria of each subject. Our data suggests that during binocular dissociation, spatial localization may be achieved by combining a reliable versional efference copy signal with a proprioceptive signal that is unreliable perhaps because it is from the wrong eye or is too noisy.

Enhanced TOA Estimation Using OFDM over Wide-Band Transmission Based on a Simulated Model

This paper presents the advantages of using a wideband spectrum adopting multi-carrier to improve targets localization within a simulated indoor environment using the Time of Arrival (TOA) technique. The study investigates the effect of using various spectrum bandwidths and a different number of carriers on localization accuracy. Also, the paper considers the influence of the transmitters’ positions in line-of-sight (LOS) and non-LOS propagation scenarios. It was found that the accuracy of the proposed method depends on the number of sub-carriers, the allocated bandwidth (BW), and the number of access points (AP). In the case of using large BW with a large number of subcarriers, the algorithm was effective to reduce localization errors compared to the conventional TOA technique. The performance degrades and becomes similar to the conventional TOA technique while using a small BW and a low number of subcarriers.

Self-supervised learning for using overhead imagery as maps in outdoor range sensor localization

Traditional approaches to outdoor vehicle localization assume a reliable, prior map is available, typically built using the same sensor suite as the on-board sensors used during localization. This work makes a different assumption. It assumes that an overhead image of the workspace is available and utilizes that as a map for use for range-based sensor localization by a vehicle. Here, range-based sensors are radars and lidars. Our motivation is simple, off-the-shelf, publicly available overhead imagery such as Google satellite images can be a ubiquitous, cheap, and powerful tool for vehicle localization when a usable prior sensor map is unavailable, inconvenient, or expensive. The challenge to be addressed is that overhead images are clearly not directly comparable to data from ground range sensors because of their starkly different modalities. We present a learned metric localization method that not only handles the modality difference, but is also cheap to train, learning in a self-supervised fashion without requiring metrically accurate ground truth. By evaluating across multiple real-world datasets, we demonstrate the robustness and versatility of our method for various sensor configurations in cross-modality localization, achieving localization errors on-par with a prior supervised approach while requiring no pixel-wise aligned ground truth for supervision at training. We pay particular attention to the use of millimeter-wave radar, which, owing to its complex interaction with the scene and its immunity to weather and lighting conditions, makes for a compelling and valuable use case.

EchoSpot

Indoor localization has played a significant role in facilitating a collection of emerging applications in the past decade. This paper presents a novel indoor localization solution via inaudible acoustic sensing, called EchoSpot, which relies on only one speaker and one microphone that are readily available on audio devices at households. We program the speaker to periodically send FMCW chirps at 18kHz-23kHz and leverage the co-located microphone to capture the reflected signals from the body and the wall for analysis. By applying the normalized cross-correlation on the transmitted and received signals, we can estimate and profile their time-of-flights (ToFs). We then eliminate the interference from device imperfection and environmental static objects, able to identify the ToFs corresponding to the direct reflection from human body. In addition, a new solution to estimate the ToF from wall reflection is designed, assisting us in spotting a human location in the two-dimensional space. We implement EchoSpot on three different types of speakers, e.g., Amazon Echo, Edifier R1280DB, and Logitech z200, and deploy them in real home environments for evaluation. Experimental results exhibit that EchoSpot achieves the mean localization errors of 4.1cm, 9.2cm, 13.1cm, 17.9cm, 22.2cm, respectively, at 1m, 2m, 3m, 4m, and 5m, comparable to results from the state-of-the-arts while maintaining favorable advantages.

Object Detection and Movement Tracking Using Tubelets and Faster RCNN Algorithm with Anchor Generation

Object detection in images and videos has become an important task in computer vision. It has been a challenging task due to misclassification and localization errors. The proposed approach explored the feasibility of automated detection and tracking of elephant intrusion along forest border areas. Due to an alarming increase in crop damages resulted from movements of elephant herds, combined with high risk of elephant extinction due to human activities, this paper looked into an efficient solution through elephant’s tracking. The convolutional neural network with transfer learning is used as the model for object classification and feature extraction. A new tracking system using automated tubelet generation and anchor generation methods in combination with faster RCNN was developed and tested on 5,482 video sequences. Real-time video taken for analysis consisted of heavily occluded objects such as trees and animals. Tubelet generated from each video sequence with intersection over union (IoU) thresholds have been effective in tracking the elephant object movement in the forest areas. The proposed work has been compared with other state-of-the-art techniques, namely, faster RCNN, YOLO v3, and HyperNet. Experimental results on the real-time dataset show that the proposed work achieves an improved performance of 73.9% in detecting and tracking of objects, which outperformed the existing approaches.

Analysis of Occlusion Effects for Map-Based Self-Localization in Urban Areas

A high-definition (HD) map provides structural information for map-based self-localization, enabling stable estimation in real environments. In urban areas, there are many obstacles, such as buses, that occlude sensor observations, resulting in self-localization errors. However, most of the existing HD map-based self-localization evaluations do not consider sudden significant errors due to obstacles. Instead, they evaluate this in terms of average error over estimated trajectories in an environment with few occlusions. This study evaluated the effects of self-localization estimation on occlusion with synthetically generated obstacles in a real environment. Various patterns of synthetic occlusion enabled the analyses of the effects of self-localization error from various angles. Our experiments showed various characteristics that locations susceptible to obstacles have. For example, we found that occlusion in intersections tends to increase self-localization errors. In addition, we analyzed the geometrical structures of a surrounding environment in high-level error cases and low-level error cases with occlusions. As a result, we suggested the concept that the real environment should have to achieve robust self-localization under occlusion conditions.

Comparison of Direct Intersection and Sonogram Methods for Acoustic Indoor Localization of Persons

We discuss two methods to detect the presence and location of a person in an acoustically small-scale room and compare the performances for a simulated person in distances between 1 and 2 m. The first method is Direct Intersection, which determines a coordinate point based on the intersection of spheroids defined by observed distances of high-intensity reverberations. The second method, Sonogram analysis, overlays all channels’ room impulse responses to generate an intensity map for the observed environment. We demonstrate that the former method has lower computational complexity that almost halves the execution time in the best observed case, but about 7 times slower in the worst case compared to the Sonogram method while using 2.4 times less memory. Both approaches yield similar mean absolute localization errors between 0.3 and 0.9 m. The Direct Intersection method performs more precise in the best case, while the Sonogram method performs more robustly.