Automatic retinal vessel segmentation has drawn significant attention in early diagnosis and treatment of many diseases, such as diabetes, retinal diseases, and coronary heart disease. However, due to vessels exhibit variations in morphology and low contrast, it is still challenging to obtain accurate segmentation results. In this paper, aiming at upgrading the accuracy and sensitivity of existing vessel segmentation methods, we propose a Multi-Scale Convolutional Neural Network with Attention Mechanisms (MSCNN-AM). For extraction of blood vessels at different scales, we introduce atrous separable convolutions with varying dilation rates, which could capture global and multi-scale vessel information better. Meanwhile, in order to reduce false-positive predictions for tiny vessel pixels, we also adopt attention mechanisms so that the proposed MSCNN-AM can pay more attention to retinal vessel pixels instead of background pixels. Because the green channel shows better vessel contrast and less noise than other channels in the RGB image, our proposed MSCNN-AM is trained and tested with green channel images only, excluding extra pre-processing and post-processing steps. The proposed method is evaluated on three public datasets, including DRIVE, STARE, and CHASE_DB1. In addition, we adopt six objective metrics to verify the performance of the MSCNN-AM, including sensitivity (Se), specificity (Sp), accuracy (Acc), F1score, an area under a receiver operating characteristic curve (AUC-ROC), and an area under precision/recall curve (AUC-PR). Experimental results indicate that our proposed method outperforms most of the existing methods with a sensitivity of 0.8342/0.8412/0.8132 and an accuracy of 0.9555/0.9658/0.9644 on DRIVE, STARE, and CHASE_DB1 separately.
Towed array shape estimation aided with non-acoustic sensors is widely used for its rather low computational complexity of solutions and rather explicit results. In addition, previous studies have emphasised that depth sensors' distribution has a dramatic influence on the accuracy of this kind of array shape estimation method. Established on the basic theory of towed array shape estimation using Kalman Filters, the approximately optimal distribution of a certain number of depth sensors over a certain number of discretised towed arrays yields the approximately best achievable performance in terms of minimum space average mean square error (AMSE), is addressed in this study. The effect of depth sensors' distribution has been discussed. Then an exact expression for the space AMSE is derived. The expression is simplified in a reasonable way considering the practical issues in order to calculate the minimum space AMSE rapidly and effectively. The performance assessments demonstrate the effectiveness of the newly proposed method.
Over the last two decades, low-frequency active sonar has become an attractive tool for underwater target detection. The reverberation to signal ratio (RSR) of transmitted waveforms is an important factor affecting the detection capability of low-frequency active sonar. Therefore, reasonable waveform design for reverberation suppression of active sonar is an important topic. Pulse trains of linear frequency-modulated (PTFM) waveforms have been proposed and manifested their good performance in suppressing reverberation. The number of sub-pulses is positively related to the reverberation to signal ratio; the lower the number of sub-pulses, the lower the reverberation to signal ratio. However, to avoid ambiguity in a Doppler measurement, the PTFM waveforms have a requirement for the number of sub-pulses to be satisfied, which prevents its reverberation suppression performance from being further improved. In this paper, we propose a coprime pulse train of linear frequency-modulated (CPTFM) waveform, which reduces the number of sub-pulses to some extent. Therefore, the ability of reverberation suppression of the CPTFM waveform can be improved. The RSR was chosen as the metric to evaluate the waveform’s ability to suppress reverberation, and the theoretical formula for the RSR of the CPTFM waveform was derived in zone A and B. With the overlap of zones A and B brought about by the decrease in the number of sub-pulses, the average RSR of zones A and B is used in this paper to evaluate the reverberation suppression ability of the waveform. The simulation experiment shows that the proposed CPTFM waveform decreases the average RSR by 7 dB and 20 dB in comparison to the reference PTFM waveform and continuous waveform (CW), which is consistent with the theoretical results by the derived formulas.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.