scholarly journals A Method for Singular Points Detection Based on Faster-RCNN

2018 ◽  
Vol 8 (10) ◽  
pp. 1853 ◽  
Author(s):  
Yonghong Liu ◽  
Baicun Zhou ◽  
Congying Han ◽  
Tiande Guo ◽  
Jin Qin
Keyword(s):  
Detection Rate ◽  
Poor Quality ◽  
Singular Points ◽  
Convolutional Network ◽  
Step Process ◽  
Detection Algorithms ◽  
Orientation Fields ◽  
Network Method ◽  
Orientation Information ◽  
Core Points

Most methods for singular points detection usually depend on the orientation fields of fingerprints, which cannot achieve reliable and accurate detection of poor quality fingerprints. In this study, a new method for fingerprint singular points detection based on Faster-RCNN (Faster Region-based Convolutional Network method) is proposed, which is a two-step process, and an orientation constraint is added in Faster-RCNN to obtain orientation information of singular points. Besides, we designed a convolutional neural network (ConvNet) for singular points detection according to the characteristics of fingerprint images and the existing works. Specifically, the proposed method could extract singular points directly from raw fingerprint images without traditional preprocessing. Experimental results demonstrate the effectiveness of the proposed method. In comparison with other detection algorithms, our method achieves 96.03% detection rate for core points and 98.33% detection rate for delta points on FVC2002 DB1 dataset while 90.75% for core points and 94.87% on NIST SD4 dataset, which outperform other algorithms.

An Image Change Detection Algorithm Based on Multi-Feature Self-Attention Fusion Mechanism UNet Network

2021 ◽  
Author(s):  
Gulnaz Alimjan ◽  
Yiliyaer Jiaermuhamaiti ◽  
Huxidan Jumahong ◽  
Shuangling Zhu ◽  
Pazilat Nurmamat
Keyword(s):  
Change Detection ◽  
Detection Algorithm ◽  
Convolutional Network ◽  
Detection Algorithms ◽  
Multiple Dimensions ◽  
Change Detection Algorithms ◽  
Image Change Detection ◽  
Control Loss ◽  
The Relationship ◽  
Entire Network

Various UNet architecture-based image change detection algorithms promote the development of image change detection, but there are still some defects. First, under the encoder–decoder framework, the low-level features are extracted many times in multiple dimensions, which generates redundant information; second, the relationship between each feature layer is not modeled so sufficiently that it cannot produce the optimal feature differentiation representation. This paper proposes a remote image change detection algorithm based on the multi-feature self-attention fusion mechanism UNet network, abbreviated as MFSAF UNet (multi-feature self-attention fusion UNet). We attempt to add multi-feature self-attention mechanism between the encoder and decoder of UNet to obtain richer context dependence and overcome the two above-mentioned restrictions. Since the capacity of convolution-based UNet network is directly proportional to network depth, and a deeper convolutional network means more training parameters, so the convolution of each layer of UNet is replaced as a separated convolution, which makes the entire network to be lighter and the model’s execution efficiency is slightly better than the traditional convolution operation. In addition to these, another innovation point of this paper is using preference to control loss function and meet the demands for different accuracies and recall rates. The simulation test results verify the validity and robustness of this approach.

DsUnet: A New Network Structure for Detection and Segmentation of Ultrasound Breast Lesions

2020 ◽  
Vol 10 (3) ◽  
pp. 661-666 ◽  
Author(s):  
Shaoguo Cui ◽  
Moyu Chen ◽  
Chang Liu
Keyword(s):  
Loss Function ◽  
Performance Metrics ◽  
Poor Quality ◽  
Breast Ultrasound ◽  
Cross Entropy ◽  
Breast Lesions ◽  
Dice Coefficient ◽  
Feature Maps ◽  
Convolutional Network ◽  
Semantic Classification

Breast cancer is one of the leading causes of death among the women worldwide. The clinical medical system urgently needs an accurate and automatic breast segmentation method in order to detect the breast ultrasound lesions. Recently, some studies show that deep learning methods based on fully convolutional network, have demonstrated a competitive performance in breast ultrasound segmentation. However, some features are missed in the Unet in case of down-sampling that results in a low segmentation accuracy. Furthermore, there is a semantic gap between the feature maps of decoder and encoder in Unet, so the simple fusion of high and low level features is not conducive to the semantic classification of pixels. In addition, the poor quality of breast ultrasound also affects the accuracy of image segmentation. To solve these problems, we propose a new end-toend network model called Dense skip Unet (DsUnet), which consists of the Unet backbone, short skip connection and deep supervision. The proposed method can effectively avoid the missing of feature information caused by down-sampling and implement the fusion of multilevel semantic information. We used a new loss function to optimize the DsUnet, which is composed of a binary cross-entropy and dice coefficient. We employed the True Positive Fraction (TPF), False Positives per image (FPs) and F -measure as performance metrics for evaluating various methods. In this paper, we adopted the UDIAT 212 dataset and the experimental results validate that our new approach achieved better performance than other existing methods in detecting and segmenting the ultrasound breast lesions. When we used the DsUnet model and new loss function (binary cross-entropy + dice coefficient), the best performance indexes can be achieved, i.e., 0.87 in TPF, 0.13 in FPs/image and 0.86 in F-measure.

scholarly journals Enhancing the accuracy of automatic eddy detection and the capability of recognizing the multi-core structures from maps of sea level anomaly

Ocean Science ◽  
2014 ◽  
Vol 10 (1) ◽  
pp. 39-48 ◽  
Author(s):  
J. Yi ◽  
Y. Du ◽  
Z. He ◽  
C. Zhou
Keyword(s):  
Sea Level ◽  
Detection Rate ◽  
Altimeter Data ◽  
Detection Accuracy ◽  
Sea Level Anomaly ◽  
Detection Algorithms ◽  
The Past ◽  
Hybrid Detection ◽  
Automated Methods ◽  
Eddy Detection

Abstract. Automated methods are important for automatically detecting mesoscale eddies in large volumes of altimeter data. While many algorithms have been proposed in the past, this paper presents a new method, called hybrid detection (HD), to enhance the eddy detection accuracy and the capability of recognizing eddy multi-core structures from maps of sea level anomaly (SLA). The HD method has integrated the criteria of the Okubo–Weiss (OW) method and the sea surface height-based (SSH-based) method, two commonly used eddy detection algorithms. Evaluation of the detection accuracy shows that the successful detection rate of HD is ~ 96.6% and the excessive detection rate is ~ 14.2%, which outperforms the OW and those methods using SLA extrema to identify eddies. The capability of recognizing multi-core structures and its significance in tracking eddy splitting or merging events have been illustrated by comparing with the detection results of different algorithms and observations in previous literature.

scholarly journals The mountaineer's method for peak detection in photoplethysmographic signals

2019 ◽  
pp. 42-50
Author(s):  
Erick Javier Argüello-Prada
Keyword(s):  
Detection Rate ◽  
Motion Artifact ◽  
Detection Algorithm ◽  
Peak Detection ◽  
Motion Artifacts ◽  
Baseline Drift ◽  
Detection Algorithms ◽  
Peak Detection Algorithm ◽  
Drift Conditions ◽  
Low Amplitude

Several efforts have been made to develop algorithms for accurate peak detection in photoplethysmographic (PPG) signals. Most of those algorithms have been specifically conceived to perform under high motion artifact and baseline drift conditions. However, little has been done regarding peak detection in low-amplitude PPG signals. In an attempt to address this issue, a simple and real-time peak detection algorithm for PPG signals was proposed. In comparison with two other well-established peak detection algorithms, the proposed method was able to achieve over than 98% sensitivity and less than 3% failed detection rate, even when the amplitude of the PPG signal dropped to 0.2 V. Still, further work is needed to improve its robustness to motion artifacts.

Screening for Diabetic Retinopathy in Primary Care: Retinal Photography Alone can Be Used Efficiently and Effectively to Exclude Those with Sight Threatening Lesions

1997 ◽  
Vol 4 (3) ◽  
pp. 174-176 ◽  
Author(s):  
P M S Evans ◽  
T S Purewal ◽  
A Hopper ◽  
H Slater ◽  
D R L Jones ◽  
...  
Keyword(s):  
Primary Care ◽  
Diabetic Retinopathy ◽  
Gold Standard ◽  
Detection Rate ◽  
False Positive Rate ◽  
Poor Quality ◽  
Background Retinopathy ◽  
Screening Performance ◽  
Positive Rate ◽  
Retinal Photography

Background— Good screening performance of retinal photography and ophthalmoscopy together in screening for diabetic retinopathy in primary care have been reported. This study reanalysed the data to evaluate the screening performance of photography alone. Methods— One thousand and ten patients screened by fundal photography and ophthalmoscopy were studied retrospectively. Fundal photographs were quality graded with poor quality pictures being excluded from the analysis. Each patient was reviewed initially by both retinal photographs and ophthalmoscopy by an ophthalmologist, the “gold standard”. Six months later the fundal photographs were reviewed and reported in a blinded manner by the ophthalmologist. Results— Two thousand and fourteen photographs were obtained, of which 162 (8%) had to be excluded because of poor quality. On review of the remaining 18S2 photographs in isolation, of 77 cases of severe retinopathy as determined by the “gold standard”, 67 had severe changes on photography—detection rate 87%. Of the 1775 cases without sight threatening retinopathy only five were judged to have sight threatening changes on photography—false positive rate 0.3%. Considering sight threatening and background retinopathy together, the detection rate was 69% (2S7 of 375) and the false positive rate 1.6% (23 of 1477). Conclusion— Good quality fundal photographs alone seem specific enough to screen for sight threatening diabetic retinopathy, but will underdetect background retinopathy.

scholarly journals Enhancing the accuracy of automatic eddy detection and the capability of recognizing the multi-core structures from maps of sea level anomaly

2013 ◽  
Vol 10 (2) ◽  
pp. 825-851 ◽  
Author(s):  
J. Yi ◽  
Y. Du ◽  
Z. He ◽  
C. Zhou
Keyword(s):  
Sea Level ◽  
Detection Rate ◽  
Altimeter Data ◽  
Detection Accuracy ◽  
Sea Level Anomaly ◽  
Detection Algorithms ◽  
Hybrid Detection ◽  
Automated Methods ◽  
Detection Evaluation ◽  
Eddy Detection

Abstract. Automated methods are important for automatically detecting mesoscale eddies in large volumes of altimeter data. While many algorithms have been proposed in the past, this paper presents a new method, called Hybrid Detection (HD), to enhance the eddy detection accuracy and the capability of recognizing eddies' multi-core structures from maps of sea level anomaly (SLA) by integrating the ideas of the Okubo–Weiss (OW) method and the sea-surface-height-based (SSH-based) method, two well-known eddy detection algorithms. Detection evaluation using an objective validation protocol shows that the HD method owns ~ 96.6% successful detection rate and ~ 14.2% excessive detection rate, which outperforms the OW method and other methods that identify eddies by SLA extrema and confirms the improvement in detection accuracy. The capability of recognizing multi-core structures and its significance in tracking eddies' splitting or merging events have been well illustrated by comparing with other detection algorithms and historical studies.

scholarly journals A semisupervised machine learning search for never-seen gravitational-wave sources

2020 ◽  
Vol 500 (4) ◽  
pp. 5408-5419
Author(s):  
Tom Marianer ◽  
Dovi Poznanski ◽  
J Xavier Prochaska
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Gravitational Wave ◽  
Outlier Detection ◽  
Predictive Models ◽  
Detection Rate ◽  
Detection Algorithms ◽  
Compact Binary ◽  
Strain Data ◽  
Cent Detection

ABSTRACT By now, tens of gravitational-wave (GW) events have been detected by the LIGO and Virgo detectors. These GWs have all been emitted by compact binary coalescence, for which we have excellent predictive models. However, there might be other sources for which we do not have reliable models. Some are expected to exist but to be very rare (e.g. supernovae), while others may be totally unanticipated. So far, no unmodelled sources have been discovered, but the lack of models makes the search for such sources much more difficult and less sensitive. We present here a search for unmodelled GW signals using semisupervised machine learning. We apply deep learning and outlier detection algorithms to labelled spectrograms of GW strain data, and then search for spectrograms with anomalous patterns in public LIGO data. We searched ${\sim}13{{\ \rm per\ cent}}$ of the coincident data from the first two observing runs. No candidates of GW signals were detected in the data analyzed. We evaluate the sensitivity of the search using simulated signals, we show that this search can detect spectrograms containing unusual or unexpected GW patterns, and we report the waveforms and amplitudes for which a $50{{\ \rm per\ cent}}$ detection rate is achieved.

Fingerprint Singular Point Detection Using Orientation Field Reliability

2011 ◽  
Vol 403-408 ◽  
pp. 4499-4506 ◽  
Author(s):  
Ravinder Kumar ◽  
Pravin Chandra ◽  
M. Hanmandlu
Keyword(s):  
Singular Point ◽  
Detection Rate ◽  
Singular Points ◽  
Identification System ◽  
Fingerprint Matching ◽  
Orientation Field ◽  
Point Detection ◽  
Field Reliability ◽  
Fingerprint Alignment ◽  
Novel Algorithm

Singular point detection is the most important step in Automatic Fingerprint Identification System (AFIS) and is used in fingerprint alignment, fingerprint matching, and particularly in classification. The computation of orientation field of a fingerprint can be verified by computing orientation field reliability. The most unreliable portion in orientation field can be the possible location of singular points. In this paper we have proposed a novel algorithm for detecting singular points using reliability of the fingerprint orientation field. Experimental results show that the proposed algorithm accurately detects singular points (core and delta) with the detection rate of 92.6 %.

Fully Convolutional Network Method of Semantic Segmentation of Class Imbalance Remote Sensing Images

2019 ◽  
Vol 39 (4) ◽  
pp. 0428004 ◽  
Author(s):  
吴止锾 Wu Zhihuan ◽  
高永明 Gao Yongming ◽  
李磊 Li Lei ◽  
薛俊诗 Xue Junshi
Keyword(s):  
Remote Sensing ◽  
Class Imbalance ◽  
Semantic Segmentation ◽  
Remote Sensing Images ◽  
Convolutional Network ◽  
Fully Convolutional Network ◽  
Network Method
Sign in / Sign up

Export Citation Format

Share Document