Irregular Segmentation Research Articles

Stroke-GFCN: ischemic stroke lesion prediction with a fully convolutional graph network.

The interpretation of image data plays a critical role during acute brain stroke diagnosis, and promptly defining the requirement of a surgical intervention will drastically impact the patient's outcome. However, determining stroke lesions purely from images can be a daunting task. Many studies proposed automatic segmentation methods for brain stroke lesions from medical images in different modalities, though heretofore results do not satisfy the requirements to be clinically reliable. We investigate the segmentation of brain stroke lesions using a geometric deep learning model that takes advantage of the intrinsic interconnected diffusion features in a set of multi-modal inputs consisting of computer tomography (CT) perfusion parameters. We propose a geometric deep learning model for the segmentation of ischemic stroke brain lesions that employs spline convolutions and unpooling/pooling operators on graphs to excerpt graph-structured features in a fully convolutional network architecture. In addition, we seek to understand the underlying principles governing the different components of our model. Accordingly, we structure the experiments in two parts: an evaluation of different architecture hyperparameters and a comparison with state-of-the-art methods. The ablation study shows that deeper layers obtain a higher Dice coefficient score (DCS) of up to 0.3654. Comparing different pooling and unpooling methods shows that the best performing unpooling method is the proportional approach, yet it often smooths the segmentation border. Unpooling achieves segmentation results more adapted to the lesion boundary corroborated with systematic lower values of Hausdorff distance. The model performs at the level of state-of-the-art models without optimized training methods, such as augmentation or patches, with a DCS of . We proposed and evaluated an end-to-end trainable fully convolutional graph network architecture using spline convolutional layers for the ischemic stroke lesion prediction. We propose a model that employs graph-based operations to predict acute stroke brain lesions from CT perfusion parameters. Our results prove the feasibility of using geometric deep learning to solve segmentation problems, and our model shows a better performance than other models evaluated. The proposed model achieves improved metric values for the DCS metric, ranging from 8.61% to 69.05%, compared with other models trained under the same conditions. Next, we compare different pooling and unpooling operations in relation to their segmentation results, and we show that the model can produce segmentation outputs that adapt to irregular segmentation boundaries when using simple heuristic unpooling operations.

MSF-Net: A Lightweight Multi-Scale Feature Fusion Network for Skin Lesion Segmentation.

Segmentation of skin lesion images facilitates the early diagnosis of melanoma. However, this remains a challenging task due to the diversity of target scales, irregular segmentation shapes, low contrast, and blurred boundaries of dermatological graphics. This paper proposes a multi-scale feature fusion network (MSF-Net) based on comprehensive attention convolutional neural network (CA-Net). We introduce the spatial attention mechanism in the convolution block through the residual connection to focus on the key regions. Meanwhile, Multi-scale Dilated Convolution Modules (MDC) and Multi-scale Feature Fusion Modules (MFF) are introduced to extract context information across scales and adaptively adjust the receptive field size of the feature map. We conducted many experiments on the public data set ISIC2018 to verify the validity of MSF-Net. The ablation experiment demonstrated the effectiveness of our three modules. The comparison experiment with the existing advanced network confirms that MSF-Net can achieve better segmentation under fewer parameters.

A video painterly stylization using semantic segmentation

ABSTRACT Most deep-learning-based style transfers methods for video extraction use features from only a single style image to perform texture synthesis. However, this does not allow users to be creative or selective. Moreover, styles are applied to both foreground objects and backgrounds. This paper presents a painterly style transfer algorithm for video based on semantic segmentation that can segment the foreground and background, enabling different stylizations. First, a fully convolutional neural network was constructed for semantic segmentation, and a GrabCut method with a dynamic bounding box was used to correct segments and refine contours and edges. Second, an enhanced motion estimation method was applied between the foreground and background objects. Third, style transfer was used to extract textures from a style image and perform texture synthesis on a content image while preserving the architecture of the content image. The proposed method not only improves the motion boundaries of optical flow but also rectifies discontinuous and irregular segmentation due to occlusion and shape deformation. Finally, the proposed method was evaluated on various videos.

The craniovertebral junction, between osseous variants and abnormalities: insight from a paleo-osteological study.

The bony components of the craniovertebral junction (CVJ) have been investigated in 172 skeletons, dug up from several archaeological sites, to define the frequency of developmental dysmorphisms, and to acquire qualitative and quantitative data about their morphology. A review of the pertinent literature is also presented. Twenty-five individuals (14.5%) exhibited at least one dysmorphism, which ranged from a condition of simple variant to a true malformation. Four individuals presented two or more anomalies at the same time (2.3% of the whole sample, 16% of the affected individuals). The most frequently observed abnormalities were: (i) the presence of a complete bony bridge in the atlas, forming a canal surrounding the vertebral artery (arcuate foramen, supertransverse foramen, and the simultaneous occurrence of arcuate foramen and supertransverse foramen); (ii) the presence of basilar processes. Basilar processes displayed a great variety in shape and dimension. They also differed with respect to their relationship with atlas and axis. The less frequently detected anomalies were: (i) complete absence of the posterior arch of C1, (ii) fusion of C2 and C3, and (iii) irregular segmentation of C2. A broad array of structural defects has been described at the CVJ. They may occur either isolated or as part of complex multisystem syndromes. Although harmless in many cases, they can notwithstanding cause severe, even life-threatening complications. When unrecognized, they may generate trouble during surgery. Hence, accurate knowledge of CVJ arrangement, including its multifarious variations, is a critical issue for radiologists, clinicians, surgeons, and chiropractors.

Increased water temperature interrupts embryonic and larval development of Indian major carp rohu Labeo rohita

Temperature is among the critical determinants of the physiology of fishes and other heterotherms throughout the life history. Temperatures outside of the optimal range distress embryonic and larval development with untoward consequences later in life. To understand the effect of high temperature on embryogenesis and organogenesis, the commercially important rohu carp (Labeo rohita) was chosen. Rohu embryos and larvae were reared at four different temperatures (30, 32, 34, and 36 °C), and indices of their hatching, development, and mortality were observed throughout early development. Embryos exposed to 30 and 32 °C showed normal development with highest rates of hatching success. Embryos at 34 °C displayed evidence of damaged zygotes, cellular deformities, damaged yolk sac coupled with shortest incubation time, and the lowest rates of hatching success. No hatching was observed at 36 °C, and these embryos displayed coagulated organs, dark yolk sac, irregular segmentation, and pustules. Larvae of rohu exposed to 34 and 36 °C showed developmental deformities (fusion in the eye, axial curvature, yolk sac ulceration, blood coagulation, tail shortening, and ulceration) and minimal survival. Our results suggest that optimal temperature of rohu embryo and larvae is 30–32 °C or lower; higher temperatures disrupt embryonic metabolism and physiology, resulting in aberrant and unsuccessful early development.

Building Earthquake Damage Analysis Using Terrestrial Laser Scanning Data

Terrestrial laser scanners (TLSs) can acquire high‐precision three‐dimensional point cloud data for earthquake‐damaged buildings. In this study, we collected TLS data in the Wenchuan earthquake zone and developed the TLS‐BSAM (terrestrial laser scanning‐based building shape analysis model) to carry out a building earthquake damage analysis. This model involves equidistance polygon array extraction, shape dispersion parameter calculations, irregular building clustering segmentation, and damage analysis. We chose 21 buildings as samples for the experiments. The results show that when using an equidistance polygon array to depict a three‐dimensional building, 0.5 m is a reasonable sampling interval for building earthquake damage analysis. Using certain characteristic parameters to carry out K‐means clustering, one can efficiently divide irregular buildings into regular blocks. Then, by weighted averages, the shape dispersion parameters can be calculated to express the damage extent to buildings. Among the shape dispersion parameters, at least the weighted average standard deviations of the tilt direction, rectangularity, compactness, and center point are suitable to reflect the damage extent. Higher values reflect more serious damage. On the basis of existing data, the weighted average standard deviations of the tilt direction and center point can be used to establish discriminant functions that can effectively distinguish the damage extent.

Geostatistical modelling of spatial dependence in area-class occurrences for improved object-based classifications of remote-sensing images

Geographical object-based image analysis (GEOBIA) is widely used for the processing of fine spatial resolution images, with increased research on contextual modelling and classification related to GEOBIA. Specifically, a previously developed object-based image classification method, known as geostatistically weighted k-NN (gk-NN) method, has shown advantages in increasing classification accuracy. The gk-NN method incorporates spatial weighting into the k-NN classifier through modelling spatial covariance of underlying area classes. However, change-of-support problem (COSP) due to different geometries of image objects is not considered therein. In this paper, we propose a method based on geostatistical de-regularization and regularization for quantifying spatial dependence in area-class occurrences and accounting for scale discrepancy in image objects and pixels. In this new modelling approach, an area-weighted (AW) distance measure is applied for modelling spatial covariance pertaining to sample image objects. The covariance model fitted with image objects sample data is de-regularized to a point-support one, so the spatial covariance over unsampled image objects can then be computed through regularization of a point-support model (RP). Unlike the previous modelling approach in the object-based gk-NN classification, whereby spatial dependence modelling is based on centroids of image objects, this method accounts for change of support and incorporates the geometry of image objects in modelling. The new modelling method was tested on three remote-sensing image subsets with different environments, using regular and irregular segmentation methods at hierarchical scales. It was confirmed that the RP method leads to largely significant increases in classification accuracies (with an average increase of 38.09% in classification accuracy with eighteen cases), compared with that by geostatistical modelling with image object centroids. The proposed method can be used for the modelling of spatial dependence in block-support data which are common in many geospatial applications.

Using convolutional neural network to identify irregular segmentation objects from very high-resolution remote sensing imagery

Convolutional neural network (CNN) has shown great success in computer vision tasks, but their application in land-use type classifications within the context of object-based image analysis has been rarely explored, especially in terms of the identification of irregular segmentation objects. Thus, a blocks-based object-based image classification (BOBIC) method was proposed to carry out end-to-end classification for segmentation objects using CNN. Specifically, BOBIC takes advantage of CNN to automatically extract complex features from the original image data, thereby avoiding the uncertainty caused by the manual extraction of features in OBIC. Additionally, OBIC compensates for the shortcomings of CNN whereby it is difficult to delineate a clear right boundary for ground objects at the pixel level. Using three high-resolution test images, the proposed BOBIC was compared with support vector machine (SVM) and random forest (RF) classifiers, and then, the effect of image blocks and mixed objects on classification accuracy was evaluated for the proposed BOBIC. Compared with conventional SVM and RF classifiers, the inclusion of CNN improved the OBIC classification performance substantially (5% to 10% increases in overall accuracy), and it also alleviated the effect derived from mixed objects.

Lossless medical image compression using geometry-adaptive partitioning and least square-based prediction.

To improve the compression rates for lossless compression of medical images, an efficient algorithm, based on irregular segmentation and region-based prediction, is proposed in this paper. Considering that the first step of a region-based compression algorithm is segmentation, this paper proposes a hybrid method by combining geometry-adaptive partitioning and quadtree partitioning to achieve adaptive irregular segmentation for medical images. Then, least square (LS)-based predictors are adaptively designed for each region (regular subblock or irregular subregion). The proposed adaptive algorithm not only exploits spatial correlation between pixels but it utilizes local structure similarity, resulting in efficient compression performance. Experimental results show that the average compression performance of the proposed algorithm is 10.48, 4.86, 3.58, and 0.10% better than that of JPEG 2000, CALIC, EDP, and JPEG-LS, respectively. Graphical abstract ᅟ.

Adaptive feature selection based on reconstruction residual and accurately located landmarks for expression-robust 3D face recognition

A novel adaptive feature selection based on reconstruction residual and accurately located landmarks for expression-robust 3D face recognition is proposed in this paper. Firstly, the novel facial coarse-to-fine landmarks localization method based on Active Shape Model and Gabor wavelets transformation is proposed to exactly and automatically locate facial landmarks in range image. Secondly, the multi-scale fusion of the pyramid local binary patterns (F-PLBP) based on the irregular segmentation associated with the located landmarks is proposed to extract the discriminative feature. Thirdly, a sparse representation-based classifier based on the adaptive feature selection (A-SRC) using the distribution of the reconstruction residual is presented to select the expression-robust feature and identify the faces. Finally, the experimental evaluation based on FRGC v2.0 indicates that the adaptive feature selection method using F-PLBP combined with the A-SRC can obtain the high recognition accuracy by performing the higher discriminative power to overcome the influence from the facial expression variations.

A new out-of-gamut determination method of image based on irregular segmentation

The out-of-gamut determination of images is very important for color gamut mapping which plays an important role in cross-media color reproduction. In this paper, aiming to achieve an accurate out-of-gamut determination of image and exploit the full potential of the reproduction device gamut, we propose a out-of-gamut determination method of images based on irregular segmentation which divides the color gamut into several parts according to the chroma and lightness of colors in CIELAB color space. First, the device color gamut is divided into the high-chroma and the low-chroma color parts, and the high-chroma parts are divided into more segments through the CIELAB a*b* plane when calculating color gamut descriptors. Then, in each segment the radiuses between the color points and the center point are calculated. For the color points located in the high-chroma parts, the color point with the biggest radius is selected as the gamut boundary descriptor, and for that in the low-chroma parts, if the corresponding outer segments are empty, the color point with the biggest radius is selected as the gamut boundary descriptor. Through the irregular segmentation of gamut, the GBDs distribute more uniformly among the color gamut surface than segment maximum method which treats all colors in the same way wherever they are located in the gamut. After that, determination of the out-of-gamut colors can be done by calculating the position relationship between the source colors and the GBDs. Additionally, GMAs would benefit from this accurate out-of-gamut determination of image. GMAs based on this out-of-gamut determination method validate the promising results.

A new out-of-gamut determination method of image based on irregular segmentation

A new out-of-gamut determination method of image based on irregular segmentation

A new device color gamut boundary description algorithm based on irregular segmentation

AbstractThe description of device color gamut is very important for color gamut mapping, which plays an important role in cross‐media color reproduction. To get a more accurate color gamut boundary, several gamut boundary description (GBD) methods have been presented in the past; however, many problems still exist in the current algorithm (e.g., accuracy and speed). In this study, we propose a new device GBD method based on irregular segmentation and evaluate it against other GBD algorithms. In the irregular segmentation method, the device color gamut is divided into the high‐chroma and the low‐chroma color parts, and the high‐chroma parts are divided into more segments through the CIELAB a*b* plane when calculating color gamut descriptors. After that, in each segment, the radii between the color points and the center point are calculated. For the color points located in the high‐chroma parts, the color point with the biggest radius is selected as the gamut boundary descriptor, and for that in the low‐chroma parts, if the corresponding outer segments are empty, the color point with the biggest radius is selected as the gamut boundary descriptor. Then, the gamut boundary descriptors are visualized in CIELAB color space using triangulation method. Finally, the irregular segmentation method is evaluated in contrast with other GBD methods using gamut mismatch index. The results show that the irregular segmentation method performs better in accuracy and has smaller variance. © 2016 Wiley Periodicals, Inc. Col Res Appl, 42, 173–181, 2017

Novel region-based image compression method based on spiking cortical model

To get the high compression ratio as well as the high-quality reconstructed image, an effective image compression scheme named irregular segmentation region coding based on spiking cortical model (ISRCS) is presented. This scheme is region-based and mainly focuses on two issues. Firstly, an appropriate segmentation algorithm is developed to partition an image into some irregular regions and tidy contours, where the crucial regions corresponding to objects are retained and a lot of tiny parts are eliminated. The irregular regions and contours are coded using different methods respectively in the next step. The other issue is the coding method of contours where an efficient and novel chain code is employed. This scheme tries to find a compromise between the quality of reconstructed images and the compression ratio. Some principles and experiments are conducted and the results show its higher performance compared with other compression technologies, in terms of higher quality of reconstructed images, higher compression ratio and less time consuming.

Congenital axis dysmorphism in a medieval skeleton

We describe here the axis dysmorphism that we observed in the skeletal remains of a human child dug up from a fifteenth century cemetery located in north-eastern Italy. This bone defect is discussed in the light of pertinent literature. We performed macroscopical examination and CT scan analysis of the axis. Axis structure was remarkably asymmetric. Whilst the left half exhibited normal morphology, the right one was smaller than normal, and its lateral articular surface showed horizontal orientation. In addition, the odontoid process appeared leftward deviated and displayed a supplementary articular-like facet situated on the right side of its surface. These findings suggest a diagnosis of unilateral irregular segmentation of atlas and axis, a rare dysmorphism dependent upon disturbances of notochordal development in early embryonic life. Likewise other malformations of the craniovertebral junction, this axis defect may alter the delicate mechanisms of upper neck movements and cause a complex series of clinical symptoms. This is an emblematic case whereby human skeletal remains may provide valuable information on the anatomical defects of craniovertebral junction.

Toxicity assessment of Amphidinium carterae, Coolia cfr. monotis and Ostreopsis cfr. ovata (Dinophyta) isolated from the northern Ionian Sea (Mediterranean Sea)

In many coastal areas the abundant proliferation of microalgae producing biotoxins determines the occurrence of Harmful Algal Blooms (HABs). Their presence in temperate waters is well documented and often associated with marine toxin-derived disease. The occurrence and toxicity of three harmful microalgae (Amphidinium carterae, Coolia cfr. monotis and Ostreopsis cfr. ovata) from the northern Ionian Sea (Mediterranean Sea) is hereby reported. The three dinoflagellates were sampled both on macroalgae and water and their morphology and occurrence were compared to those of other Mediterranean sites. The toxicity of the three cultured strains was tested by Artemia salina and hemolysis tests and their effects on the first stages of the sea urchin development was also evaluated. The contemporary presence of the three species inhibited the in vitro sea urchin embryonic development. But this action could be ascribed to the sole Ostreopsis as the addition of the single species to the sea urchins embryos evidenced no effects in presence of Amphidinium or Coolia cells, and an irregular segmentation in presence of Ostreopsis. In particular, this latter species exerted a cytotoxic effect in a dose-dependent manner, with a production of deformed embryos even at very low cell concentration (42 cells mL−1). Nevertheless, when algal cell lysate was added, some effects on the sea urchin development was detected for each dinoflagellate, and also in this case Ostreopsis has proved to be the most toxic species. However, the lysate of Amphidinium and Ostreopsis strongly affects the A. salina nauplii vitality, while the hemolytic activity was very low for Amphidinium and Coolia lysate and very strong for Ostreopsis.Our results highlight the importance to monitoring the presence of these dinoflagellates whose effects may also be reflected on the early life stages of marine organisms, especially those species that are important from both an ecological and economic point of view, as the sea urchins are.

Content-based irregularly shaped macroblock partition for inter frame prediction in video coding

Although rate–distortion adaptive, variable block-size motion compensation has significantly contributed to the performance improvement in compression efficiency of H.264/AVC over the other existing video coding standards, the shape of the block for motion compensation prediction (MCP) in H.264/AVC is still limited to the regular rectangle. To further improve the accuracy of MCP, a novel content-adaptive macroblock partition method for inter frame prediction is proposed in this paper, where irregularly shaped partitions are introduced to MCP. The proposed method irregularly segments the current coding macroblock (MB) by referring to the matching reference region (MRR) in the reference frames, which is obtained relying on the inherent structural similarity of the video contents among successive frames and can be traced with the relative displacement vector between the current MB and the MRR. Consequently, the irregular partition of the current MB is content adaptive and the partition information is implicitly indicated by the MRR, without resorting to an explicit description of the shape information of the irregular segmentation. Experimental results demonstrate that the proposed method outperforms the MCP based on the regular rectangle blocks in terms of coding efficiency, by providing more precise motion estimation for macroblocks at little cost of extra side information.

2D Log-Gabor Wavelet Based Action Recognition

The frequency response of log-Gabor function matches well the frequency response of primate visual neurons. In this letter, motion-salient regions are extracted based on the 2D log-Gabor wavelet transform of the spatio-temporal form of actions. A supervised classification technique is then used to classify the actions. The proposed method is robust to the irregular segmentation of actors. Moreover, the 2D log-Gabor wavelet permits more compact representation of actions than the recent neurobiological models using Gabor wavelet.

Effects of space environment on embryonic growth up to hatching of salamander eggs fertilized and developed during orbital flights.

In vertebrates, only few experiments have been performed in microgravity to study the embryonic development from fertilization. To date, these concern only amphibian and fish. We report here a study on the embryonic development of Pleurodeles waltl (urodele amphibian) eggs oviposited in microgravity. The experiment was performed twice on board the Mir space station and the data obtained included video recording and morphological, histological and immunocytological analyses. The data confirm that the microgravity conditions have effects during the embryonic period, particularly during cleavage and neurulation, inducing irregular segmentation and abnormal closure of the neural tube. Moreover, we observed several abnormalities hither to undescribed corresponding to cortical cytoplasm movements, a decrease of cell adhesion and a loss of cells. These abnormalities were temporary and subsequently reversible. The young larvae that hatched during the flight displayed normal morphology and swimming behavior after landing. The results obtained in the urodele Pleurodeles waltl are in accordance with those observed earlier in the anuran Xenopus laevis and in the fish Oryzias latipes.

Consequences of somite manipulation on the pattern of dorsal root ganglion development.

We have investigated dorsal root ganglion formation, in the avian embryo, as a function of the composition of the paraxial somitic mesoderm. Three or four contiguous young somites were unilaterally removed from chick embryos and replaced by multiple cranial or caudal half-somites from quail embryos. Migration of neural crest cells and formation of DRG were subsequently visualized both by the HNK-1 antibody and the Feulgen nuclear stain. At advanced migratory stages (as defined by Teillet et al. Devl Biol. 120, 329-347 1987), neural crest cells apposed to the dorsolateral faces of the neural tube were distributed in a continuous, nonsegmented pattern that was indistinguishable on unoperated sides and on sides into which either half of the somites had been grafted. In contrast, ventrolaterally, neural crest cells were distributed segmentally close to the neural tube and within the cranial part of each normal sclerotome, whereas they displayed a nonsegmental distribution when the graft involved multiple cranial half-somites or were virtually absent when multiple caudal half-somites had been implanted. In spite of the identical dorsal distribution of neural crest cells in all embryos, profound differences in the size and segmentation of DRG were observed during gangliogenesis (E4-9) according to the type of graft that had been performed. Thus when the implant consisted of compound cranial half-somites, giant, coalesced ganglia developed, encompassing the entire length of the graft. On the other hand, very small, dorsally located ganglia with irregular segmentation were seen at the level corresponding to the graft of multiple caudal half-somites. We conclude that normal morphogenesis of dorsal root ganglia depends upon the craniocaudal integrity of the somites.