Proposed Segmentation Scheme Research Articles

Aorta Segmentation in 3D CT Images by Combining Image Processing and Machine Learning Techniques.

Aorta segmentation is extremely useful in clinical practice, allowing the diagnosis of numerous pathologies, such as dissections, aneurysms and occlusive disease. In such cases, image segmentation is prerequisite for applying diagnostic algorithms, which in turn allow the prediction of possible complications and enable risk assessment, which is crucial in saving lives. The aim of this paper is to present a novel fully automatic 3D segmentation method, which combines basic image processing techniques and more advanced machine learning algorithms, for detecting and modelling the aorta in 3D CT imaging data. An initial intensity threshold-based segmentation procedure is followed by a classification-based segmentation approach, based on a Markov Random Field network. The result of the proposed two-stage segmentation process is modelled and visualized. The proposed methodology was applied to 16 3D CT data sets and the extracted aortic segments were reconstructed as 3D models. The performance of segmentation was evaluated both qualitatively and quantitatively against other commonly used segmentation techniques, in terms of the accuracy achieved, compared to the actual aorta, which was defined manually by experts. The proposed methodology achieved superior segmentation performance, compared to all compared segmentation techniques, in terms of the accuracy of the extracted 3D aortic model. Therefore, the proposed segmentation scheme could be used in clinical practice, such as in treatment planning and assessment, as it can speed up the evaluation of the medical imaging data, which is commonly a lengthy and tedious process.

A hybrid machine-learning approach for analysis of methane hydrate formation dynamics in porous media with synchrotron CT imaging.

Fast multi-phase processes in methane hydrate bearing samples pose a challenge for quantitative micro-computed tomography study and experiment steering due to complex tomographic data analysis involving time-consuming segmentation procedures. This is because of the sample's multi-scale structure, which changes over time, low contrast between solid and fluid materials, and the large amount of data acquired during dynamic processes. Here, a hybrid approach is proposed for the automatic segmentation of tomographic data from time-resolved imaging of methane gas-hydrate formation in sandy granular media, which includes a deep-learning 3D U-Net model. To prepare a training dataset for the 3D U-Net, a technique to automate data labeling based on sample-specific information about the mineral matrix immobility and occasional fluid movement in pores is proposed. Automatic segmentation allowed for studying properties of the hydrate growth in pores, as well as dynamic processes such as incremental flow and redistribution of pore brine. Results of the quantitative analysis showed that for typical gas-hydrate stability parameters (100 bar methane pressure, 7°C temperature) the rate of formation is slow (less than 1% per hour), after which the surface area of contact between brine and gas increases, resulting in faster formation (2.5% per hour). Hydrate growth reaches the saturation point after 11 h of the experiment. Finally, the efficacy of the proposed segmentation scheme in on-the-fly automatic data analysis and experiment steering with zooming to regions of interest is demonstrated.

Remote Sensing Image Segmentation of Mariculture Cage Using Ensemble Learning Strategy

In harbour areas, the irrational layout and high density of mariculture cages can lead to a dramatic deterioration of the culture’s ecology. Therefore, it is important to analyze and regulate the distribution of cages using intelligent analysis based on deep learning. We propose a remote sensing image segmentation method based on the Swin Transformer and ensemble learning strategy. Firstly, we collect multiple remote sensing images of cages and annotate them, while using data expansion techniques to construct a remote sensing image dataset of mariculture cages. Secondly, the Swin Transformer is used as the backbone network to extract the remote sensing image features of cages. A strategy of alternating the local attention module and the global attention module is used for model training, which has the benefit of reducing the attention computation while exchanging global information. Then, the ensemble learning strategy is used to improve the accuracy of remote sensing cage segmentation. We carry out quantitative and qualitative analyses of remote sensing image segmentation of cages at the ports of Li’an, Xincun and Potou in Hainan Province, China. The results show that our proposed segmentation scheme has significant performance improvement compared to other models. In particular, the mIoU reaches 82.34% and pixel accuracy reaches 99.71%.

IRS-Enabled Beam-Space Channel

The intelligent reflecting surface (IRS) is emphasized as a controlled scattering cluster. To this end, scatterers and traveling paths of multipath components are classified to build a new channel model. Unlike the conventional modeling, where the channels between system units are modeled independently, the new model considers the channel as a whole and decomposes it based on the traveling paths. The model shows clearly how IRS, in the beam-space context, converts the channel from a problem into a design element. After investigating IRS as a scattering cluster, based on a proposed segmentation scheme, the beamforming problem is considered with a focus on first-order reflections. Passive beamforming at IRS is shown to have two tiers; at the scatterer and antenna levels. A segment-activation scheme is proposed to maximize the received signal power, where the number of transmitting antenna elements to be used is given as a function of IRS positioning and beamforming at the receiver. The results show that while using more transmitting antenna elements to get narrower beams is possible, using fewer elements can give better performance, especially for larger IRS at close distances. The developed model also proves useful in addressing emerging issues in massive MIMO communication, namely, stationarity and spherical wavefronts.

Secure Cancelable Template Based on Double Random Phase Encoding and Entropy Segmentation

In this paper, a proposed cancellable biometric scheme is based on multiple biometric image identifiers, Arnold’s cat map and double random phase encoding (DRPE) to obtain cancellable biometric templates. The proposed segmentation scheme that is used to select the region of interest for generating cancelable templates is based on chaos entropy low correlation statistical metrics. The objective of segmentation is to reduce the computational cost and reliability of template creation. The left and right biometric (iris, fingerprint, palm print and face) are divided into non-overlapping blocks of the same dimensions. To define the region of interest (ROI), we select the block with the highest entropy. To shorten the registration process time and achieve a high level of security, we select 25% of the image volume of the biometric data. In addition, the low-cost security requirement lies in the use of selective encryption (SE) technology. The step of selecting the maximum entropy is executed on all biometric blocks. The maximum right and left multi-biometric blocks are arranged in descending order from the entropy perspective and select 50% of each biometric couple and store the single matrix. The obtained matrix is scrambled with a certain number of iterations using Arnold’s Cat Map (ACM).The obtained scrambled matrix is encrypted with the DRPE to generate the cancellable biometric templates, which are further concatenated. The simulation results display better performance of the suggested cancellable biometric system in noise scenarios using the area under the receiver operating characteristic (AROC). The strength of the suggested technique is examined with correlation, irregular deviation, maximum difference and maximum deviation. The recommended proposed approach shows that the ability to distinguish the authentic and imposter biometrics of user seven in different levels of the noise environment.

Hybrid Segmentation Scheme for Skin Features Extraction Using Dermoscopy Images

Objective and quantitative assessment of skin conditions is essential for cosmeceutical studies and research on skin aging and skin regeneration. Various handcraft-based image processing methods have been proposed to evaluate skin conditions objectively, but they have unavoidable disadvantages when used to analyze skin features accurately. This study proposes a hybrid segmentation scheme consisting of Deeplab v3+ with an Inception-ResNet-v2 backbone, LightGBM, and morphological processing (MP) to overcome the shortcomings of handcraft-based approaches. First, we apply Deeplab v3+ with an Inception-ResNet-v2 backbone for pixel segmentation of skin wrinkles and cells. Then, LightGBM and MP are used to enhance the pixel segmentation quality. Finally, we determine several skin features based on the results of wrinkle and cell segmentation. Our proposed segmentation scheme achieved a mean accuracy of 0.854, mean of intersection over union of 0.749, and mean boundary F1 score of 0.852, which achieved 1.1%, 6.7%, and 14.8% improvement over the panoptic-based semantic segmentation method, respectively.

A Novel Parallel DTC Segmentation Scheme for Fractional-N Digital PLLs

In this paper, we propose a new parallel segmentation scheme for the digital/time converter (DTC) which is employed in fractional-N digital phase-locked loops (PLLs) to cancel out the quantization error induced by the digital DS modulator. The proposed parallel scheme removes one redundant least-mean square (LMS) gain in compared with the conventional parallel one. Therefore, the design of the system becomes less complicated while guaranteeing a fast convergence speed of the LMS gains and a short DTC time range. The effectiveness of the proposed segmentation scheme is demonstrated via simulations of a digital PLL built at behavioral level and compared to the conventional segmentation schemes.

Power-Law Processor Over Segmentation for Variable Length Transients Detection

In this letter, we aim to apply Nuttall's power-law processor, for which available data are raised to a power prior to summation, to the detection of variable length transient signals. The available data are often couched in frequency domain, where aggregation of transient signal energy into relatively small number of frequency bins is achieved by matching the DFT size with the transient length. However, without prior knowledge of the transient duration, performance degrades due to mismatch and the spread of transient energy over frequency bins as a result. A simple segmentation scheme is proposed to deal with the mismatch problem by exploiting robustness of power-law processor and contiguity of transients in time domain. Performance of the proposed segmentation scheme is evaluated, and is compared with two time domain detectors with and without the prior knowledge of transient length.

Patch Based Deep Local Feature Learning and Self Similarity Multi Level Clustering for Neonatal Brain Segmentation in MR Images

The main purpose of this work is to develop a new scheme to profoundly retrieving features to perform the process of identifying a brain regions from the MR neonatal brain image. First the input MR neonatal brain image is denoised by using the Modified Fuzzy Adaptive Non Local Mean Filter (FANLMF) and then the contrast of the image is enhanced using the Adaptive Average Intensity Based Histogram Equalization (AAIHE). After pre-processing the input MR image, the next step is to retrieve the features of a similar image. To capture the features from the pre-processed image, this project offers a new technique for retrieving features called Patch Based Deep Local Feature Learning (PBDLFL). After retrieving the deep features, the next step is to divide the brain regions based on these retrieved features. To implement this process, the supervised segmentation scheme is employed. Among several supervised segmentation scheme this works employs proposed approach named Self Similarity Multi Level Clustering (SSMLC). Finally, the retrieved features are given as an input to these SSMLC approach for separating the regions of the brain. To understand the effectiveness of the proposed deep feature retrieval and proposed segmentation scheme, four performance metrics are employed namely, Dice Similarity Coefficient (DSC), Positive Predictive Value (PPV), Jaccard Index (JI) and Sensitivity (SEN). The experimental results show that the new PBDLFL and SSMLC perform better than other existing approaches.

Cuttlefish Algorithm-Based Multilevel 3-D Otsu Function for Color Image Segmentation

To overcome the shortcomings of 1-D and 2-D Otsu's thresholding methods, a 3-D Otsu method has been introduced. While yielding satisfactory segmentation results for images with a low signal-to-noise ratio (SNR) and poor contrast, it has the downside of high computational complexity. In this paper, the cuttlefish algorithm (CFA)-based 3-D Otsu thresholding method is proposed to pace up the conventional 3-D Otsu thresholding for color image segmentation. In order to decrease the effects of noises and weak edges, an optimally selected multilevel 3-D Otsu image thresholding technique is brought into the proposed segmentation scheme. The CFA is a newly developed stochastic meta-heuristic optimization algorithm based on observing, mimicking, and modeling the camouflaging feature of cuttlefish. It is used to simplify the problem of exhaustive search for the optimal threshold vector in 3-D space. Experimental results, when compared to 1-D Otsu, 1-D Otsu-Cuckoo search (CS) algorithm, 1-D Otsu-lightning search algorithm (LSA), 1-D Otsu-CFA, conventional 3-D Otsu, 3-D Otsu-CS, and 3-D Otsu-LSA, indicate that the proposed algorithm CFA-based 3-D Otsu thresholding is superior to all the other multilevel thresholding algorithms. The proposed 3-D-CFA method produces promising segmentation results from the objective and subjective aspects.

Fully Convolutional Networks for Monocular Retinal Depth Estimation and Optic Disc-Cup Segmentation.

Glaucoma is a serious ocular disorder for which the screening and diagnosis are carried out by the examination of the optic nerve head (ONH). The color fundus image (CFI) is the most common modality used for ocular screening. In CFI, the central region which is the optic disc and the optic cup region within the disc are examined to determine one of the important cues for glaucoma diagnosis called the optic cup-to-disc ratio (CDR). CDR calculation requires accurate segmentation of optic disc and cup. Another important cue for glaucoma progression is the variation of depth in ONH region. In this paper, we first propose a deep learning framework to estimate depth from a single fundus image. For the case of monocular retinal depth estimation, we are also plagued by the labeled data insufficiency. To overcome this problem we adopt the technique of pretraining the deep network where, instead of using a denoising autoencoder, we propose a new pretraining scheme called pseudo-depth reconstruction, which serves as a proxy task for retinal depth estimation. Empirically, we show pseudo-depth reconstruction to be a better proxy task than denoising. Our results outperform the existing techniques for depth estimation on the INSPIRE dataset. To extend the use of depth map for optic disc and cup segmentation, we propose a novel fully convolutional guided network, where, along with the color fundus image the network uses the depth map as a guide. We propose a convolutional block called multimodal feature extraction block to extract and fuse the features of the color image and the guide image. We extensively evaluate the proposed segmentation scheme on three datasets- ORIGA, RIMONEr3, and DRISHTI-GS. The performance of the method is comparable and in many cases, outperforms the most recent state of the art.

A Segmentation Scheme for CLF Dynamic Movement Control Applied to Robotic Handwriting

Control Lyapunov Function - Dynamic Movement is a recently proposed imitation learning approach for controlling point-to-point movements. In this work some practical issues arising with this approach have been investigated by studying a particular application: robotic handwriting of German letters. One difficulty in this context is the handling of crossings. Here a novel segmentation scheme is proposed to eliminate crossings. Using a segmentation framework for splitting up the movement into a number of partial movements has additional advantages: It reduces the computational time for calculating the dynamic system that is required in this approach, as well as the time for calculating a Lyapunov function because of the simpler parametrization. This proposed segmentation scheme has been validated experimentally on a CRS A465 robotic manipulator.

Coexistence of Smart Utility Networks and WLANs in Smart Grid Systems

This paper aims to analyze coexistence performance and study the impact of WLAN interferences on SUNs. To quantitatively reveal packet error rate (PER) performance of SUNs under multiple WLAN interferers, a segmentation scheme for desired data packet with multiple interferers is developed, based on which a packet collision model in a multi-interferer environment is proposed. The new packet collision model captures interference scenarios in every time segment of a desired packet using the proposed segmentation scheme. It paves a way to make a precise estimation of the PER of SUNs under multiple WLAN interferers. Furthermore, we design a channel agility scheme for coexistence of SUNs and WLANs via improving the existing ones, and analyze its performance in interference identification and avoidance. The proposed scheme performs interference detection for every data packet, and thus, an SUN is able to avoid WLANs interferences in a timely manner.

Brain-inspired speech segmentation for automatic speech recognition using the speech envelope as a temporal reference.

Speech segmentation is a crucial step in automatic speech recognition because additional speech analyses are performed for each framed speech segment. Conventional segmentation techniques primarily segment speech using a fixed frame size for computational simplicity. However, this approach is insufficient for capturing the quasi-regular structure of speech, which causes substantial recognition failure in noisy environments. How does the brain handle quasi-regular structured speech and maintain high recognition performance under any circumstance? Recent neurophysiological studies have suggested that the phase of neuronal oscillations in the auditory cortex contributes to accurate speech recognition by guiding speech segmentation into smaller units at different timescales. A phase-locked relationship between neuronal oscillation and the speech envelope has recently been obtained, which suggests that the speech envelope provides a foundation for multi-timescale speech segmental information. In this study, we quantitatively investigated the role of the speech envelope as a potential temporal reference to segment speech using its instantaneous phase information. We evaluated the proposed approach by the achieved information gain and recognition performance in various noisy environments. The results indicate that the proposed segmentation scheme not only extracts more information from speech but also provides greater robustness in a recognition test.

Multi-scale patch and multi-modality atlases for whole heart segmentation of MRI.

A whole heart segmentation (WHS) method is presented for cardiac MRI. This segmentation method employs multi-modality atlases from MRI and CT and adopts a new label fusion algorithm which is based on the proposed multi-scale patch (MSP) strategy and a new global atlas ranking scheme. MSP, developed from the scale-space theory, uses the information of multi-scale images and provides different levels of the structural information of images for multi-level local atlas ranking. Both the local and global atlas ranking steps use the information theoretic measures to compute the similarity between the target image and the atlases from multiple modalities. The proposed segmentation scheme was evaluated on a set of data involving 20 cardiac MRI and 20 CT images. Our proposed algorithm demonstrated a promising performance, yielding a mean WHS Dice score of 0.899 ± 0.0340, Jaccard index of 0.818 ± 0.0549, and surface distance error of 1.09 ± 1.11 mm for the 20 MRI data. The average runtime for the proposed label fusion was 12.58 min.

Learning automata for image segmentation

A method is proposed seeking thresholds for segmentation of grayscale images. The normalized image histogram is modeled as a Gaussian mixture, and the parameters associated with the Gaussian components are estimated iteratively with a set of learning automata. To reduce the parameter search space, the number of major components in the image and their associated parameter ranges are first specified using some desired properties of the Gaussian distribution. Thresholds are chosen based on the Gaussian parameter estimates after convergence. Illustrative examples are provided to demonstrate the learning process and the effectiveness of the proposed segmentation scheme.

Human habenula segmentation using myelin content

The habenula consists of a pair of small epithalamic nuclei located adjacent to the dorsomedial thalamus. Despite increasing interest in imaging the habenula due to its critical role in mediating subcortical reward circuitry, in vivo neuroimaging research targeting the human habenula has been limited by its small size and low anatomical contrast. In this work, we have developed an objective semi-automated habenula segmentation scheme consisting of histogram-based thresholding, region growing, geometric constraints, and partial volume estimation steps. This segmentation scheme was designed around in vivo 3T myelin-sensitive images, generated by taking the ratio of high-resolution T1w over T2w images. Due to the high myelin content of the habenula, the contrast-to-noise ratio with the thalamus in the in vivo 3T myelin-sensitive images was significantly higher than the T1w or T2w images alone. In addition, in vivo 7T myelin-sensitive images (T1w over T2*w ratio images) and ex vivo proton density-weighted images, along with histological evidence from the literature, strongly corroborated the in vivo 3T habenula myelin contrast used in the proposed segmentation scheme. The proposed segmentation scheme represents a step toward a scalable approach for objective segmentation of the habenula suitable for both morphological evaluation and habenula seed region selection in functional and diffusion MRI applications.

In Situ 3D Segmentation of Individual Plant Leaves Using a RGB-D Camera for Agricultural Automation.

In this paper, we present a challenging task of 3D segmentation of individual plant leaves from occlusions in the complicated natural scene. Depth data of plant leaves is introduced to improve the robustness of plant leaf segmentation. The low cost RGB-D camera is utilized to capture depth and color image in fields. Mean shift clustering is applied to segment plant leaves in depth image. Plant leaves are extracted from the natural background by examining vegetation of the candidate segments produced by mean shift. Subsequently, individual leaves are segmented from occlusions by active contour models. Automatic initialization of the active contour models is implemented by calculating the center of divergence from the gradient vector field of depth image. The proposed segmentation scheme is tested through experiments under greenhouse conditions. The overall segmentation rate is 87.97% while segmentation rates for single and occluded leaves are 92.10% and 86.67%, respectively. Approximately half of the experimental results show segmentation rates of individual leaves higher than 90%. Nevertheless, the proposed method is able to segment individual leaves from heavy occlusions.

Smartphone based visible iris recognition using deep sparse filtering

Good biometric performance of iris recognition motivates it to be used for many large scale security and access control applications. Recent works have identified visible spectrum iris recognition as a viable option with considerable performance. Key advantages of visible spectrum iris recognition include the possibility of iris imaging in on-the-move and at-a-distance scenarios as compared to fixed range imaging in near-infra-red light. The unconstrained iris imaging captures the images with largely varying radius of iris and pupil. In this work, we propose a new segmentation scheme and adapt it to smartphone based visible iris images for approximating the radius of the iris to achieve robust segmentation. The proposed technique has shown the improved segmentation accuracy up to 85% with standard OSIRIS v4.1. This work also proposes a new feature extraction method based on deep sparse filtering to obtain robust features for unconstrained iris images. To evaluate the proposed segmentation scheme and feature extraction scheme, we employ a publicly available database and also compose a new iris image database. The newly composed iris image database (VSSIRIS) is acquired using two different smartphones – iPhone 5S and Nokia Lumia 1020 under mixed illumination with unconstrained conditions in visible spectrum. The biometric performance is benchmarked based on the equal error rate (EER) obtained from various state-of-art schemes and proposed feature extraction scheme. An impressive EER of 1.62% is obtained on our VSSIRIS database and an average gain of around 2% in EER is obtained on the public database as compared to the well-known state-of-art schemes.

Trabecular Bone Image Segmentation Using Wavelet and Marker-Controlled Watershed Transformation

This paper presents a new strategy for the segmentation of trabecular bone image. This kind of image is acquired with microcomputed tomography (micro-CT) to assess bone microarchitecture based chiefly on bone mineral density (BMD) measurements to improve fracture risk prediction. Disease osteoporosis can be predicted from features of CT image where a bone region may consist of several disjoint pieces. It relies on a multiresolution representation of the image by the wavelet transform to compute the multiscale morphological gradient. The coefficients of detail found at the different scales are used to determine the markers and homogeneous regions that are extracted with the watershed algorithm. The method reduces the tendency of the watershed algorithm to oversegment and results in closed homogeneous regions. The performance of the proposed segmentation scheme is presented via experimental results obtained with a broad series of images.