Quantitative Analysis of Patch-Based Fully Convolutional Neural Networks for Tissue Segmentation on Brain Magnetic Resonance Imaging

Abdominal image segmentation is a challenging, yet important clinical problem. Variations in body size, position, and relative organ positions greatly complicate the segmentation process. Historically, multi-atlas methods have achieved leading results across imaging modalities and anatomical targets. However, deep learning is rapidly overtaking classical approaches for image segmentation. Recently, Zhou et al. showed that fully convolutional networks produce excellent results in abdominal organ segmentation of computed tomography (CT) scans. Yet, deep learning approaches have not been applied to whole abdomen magnetic resonance imaging (MRI) segmentation. Herein, we evaluate the applicability of an existing fully convolutional neural network (FCNN) designed for CT imaging to segment abdominal organs on T2 weighted (T2w) MRI's with two examples. In the primary example, we compare a classical multi-atlas approach with FCNN on forty-five T2w MRI's acquired from splenomegaly patients with five organs labeled (liver, spleen, left kidney, right kidney, and stomach). Thirty-six images were used for training while nine were used for testing. The FCNN resulted in a Dice similarity coefficient (DSC) of 0.930 in spleens, 0.730 in left kidneys, 0.780 in right kidneys, 0.913 in livers, and 0.556 in stomachs. The performance measures for livers, spleens, right kidneys, and stomachs were significantly better than multi-atlas (p < 0.05, Wilcoxon rank-sum test). In a secondary example, we compare the multi-atlas approach with FCNN on 138 distinct T2w MRI's with manually labeled pancreases (one label). On the pancreas dataset, the FCNN resulted in a median DSC of 0.691 in pancreases versus 0.287 for multi-atlas. The results are highly promising given relatively limited training data and without specific training of the FCNN model and illustrate the potential of deep learning approaches to transcend imaging modalities.

A Quantum-Inspired Self-Supervised Network model for automatic segmentation of brain MR images

In this work, we propose a novel learning-based segmentation technique for delineating liver volumes in magnetic resonance images. The method utilizes the shape prior of the liver for improved accuracy. Instead of labeling the tissue via binary classification, our method completes the segmentation by deforming a label template of the liver average shape based on the learned image features. The average shape of the liver we used is estimated from a large set of expert-labeled computed tomography images. A fully convolutional neural network (FCN) is trained to maximize the overlap between the deformed liver label template and the ground truth segmentation. The proposed method is validated with 51 T2-weighted liver image volumes and achieves an average Dice coefficient of 95.2% with a mean Hausdorff distance of 20.0 mm. Compared to the results obtained with a standard FCN-based method, a three-fold improvement of the Hausdorff distance is observed, indicating the substantial gains achieved by incorporating the shape prior.

The brain is considered a main important part of human body as it participate to control all the activities of human. Under some unhealthy conditions the abnormal and uncontrolled growth of tissue may take place which is unnatural process. This kind of process taking place in human brain is referred as “brain tumor”. The brain tumor may be of cancerous (malignant) or may be of non-cancerous (Benign) and its detection is a very much important to reduced the death rate of humans suffering from tumor. The detection of brain tumor can be performed by using various image processing techniques like brain magnetic resonance imaging (MRI), Computed Tomography (CT), Electroencephalography (EEG), Positron Emission Tomography (PET), Magneto Encephalography (MEG) etc. Among these techniques the brain MRI is widely adopted in the world due to its significant features. But to detect tumor in the brain MRI, the processes like preprocessing, segmentation and classification are must. This paper deals with the various aspects of the brain tumor detection. The paper discusses the significant researches which are meant for brain tumor detection through MRI quality enhancement, MRI segmentation and MRI classification (as cancerous or not). The outcome of the paper discussed that there were no proper researches are performed to detect the brain tumor, focused on good quality of image, less researches on segmentation and classification of brain tumor. Finally the paper gives the future research scope to overcome the addressed research gap.

A local Zernike moment-based unbiased nonlocal means fuzzy C-Means algorithm for segmentation of brain magnetic resonance images

Bone segmentation in magnetic resonance imaging (MRI) is crucial for clinical and research applications, including diagnosis, surgical planning, and treatment monitoring. However, it remains challenging due to anatomical variability and complex bone morphology. Manual segmentation is time-consuming and operator-dependent, fostering interest in automated methods. This study proposes an automated segmentation method based on a 3D U-Net convolutional neural network to segment the femur, tibia, and patella from low-field MRI scans. Low-field MRI offers advantages in cost, patient comfort, and accessibility but presents challenges related to lower signal quality. Our method achieved a Dice Similarity Coefficient (DSC) of 0.9838, Intersection over Union (IoU) of 0.9682, and Average Hausdorff Distance (AHD) of 0.0223, with an inference time of approximately 3.96 s per volume on a GPU. Although post-processing had minimal impact on metrics, it significantly enhanced the visual smoothness of bone surfaces, which is crucial for clinical use. The final segmentations enabled the creation of clean, 3D-printable bone models, beneficial for preoperative planning. These results demonstrate that the model achieves accurate segmentation with a high degree of overlap compared to manually segmented reference data. This accuracy results from meticulous fine-tuning of the network, along with the application of advanced data augmentation and post-processing techniques.

Brain tumor segmentation in Magnetic Resonance Imaging (MRI) scans provides vital information to radiologists in the diagnosis and staging of disease. However, these MRI scans are often corrupted with noise during its acquisition. Traditional approaches to this problem employ denoising which leads, in general, to edge smoothing and development of artifacts in MRI slices, thereby affecting tumor segmentation performance. In this paper, we employed graph signal processing (GSP) theory to first segment tumor core in each MRI slice using graph Laplacian followed by edge-aware denoising which is performed in synergy. The paper aims to present a novel technique to tackle the two problems of segmentation and denoising both under the GSP framework. The experimental results demonstrated on simulated and clinical brain MRI datasets, show highly competitive performance both in terms of tumor core segmentation under Dice and Sensitivity measures, and in terms of edge-aware denoising under PSNR and SSIM measures.

Surgical robots are gaining increasing popularity because of their capability to improve the precision of pedicle screw placement. However, current surgical robots rely on unimodal computed tomography (CT) images as baseline images, limiting their visualization to vertebral bone structures and excluding soft tissue structures such as intervertebral discs and nerves. This inherent limitation significantly restricts the applicability of surgical robots. To address this issue and further enhance the safety and accuracy of robot-assisted pedicle screw placement, this study will develop a software system for surgical robots based on multimodal image fusion. Such a system can extend the application range of surgical robots, such as surgical channel establishment, nerve decompression, and other related operations. Initially, imaging data of the patients included in the study are collected. Professional workstations are employed to establish, train, validate, and optimize algorithms for vertebral bone segmentation in CT and magnetic resonance (MR) images, intervertebral disc segmentation in MR images, nerve segmentation in MR images, and registration fusion of CT and MR images. Subsequently, a spine application model containing independent modules for vertebrae, intervertebral discs, and nerves is constructed, and a software system for surgical robots based on multimodal image fusion is designed. Finally, the software system is clinically validated. We will develop a software system based on multimodal image fusion for surgical robots, which can be applied to surgical access establishment, nerve decompression, and other operations not only for robot-assisted nail placement. The development of this software system is important. First, it can improve the accuracy of pedicle screw placement, percutaneous vertebroplasty, percutaneous kyphoplasty, and other surgeries. Second, it can reduce the number of fluoroscopies, shorten the operation time, and reduce surgical complications. In addition, it would be helpful to expand the application range of surgical robots by providing key imaging data for surgical robots to realize surgical channel establishment, nerve decompression, and other operations.

Accurate and fast segmentation of magnetic resonance (MR) images with volumetric brain formations, such as glioblastomas and meningiomas, helps plan surgical and radiation treatment, increase the safety and radicality of surgical intervention, which in turn allows to increase the overall life expectancy of patients and the disease-free period. It is required for the development and evaluation of the effectiveness of a deep convolutional neural network with transmission connections for automatic segmentation of volumetric brain formations (meningiomas and glioblastomas) on postoperative MR images, as well as analysis of its accuracy in comparison with manual expert segmentation and existing methods. Material and methods . The paper considers the creation of an architecture inspired by the SegResNet model, training on BraTS2024-GLI and BraTS2024-MEN-RT, describes a method for compiling a training sample that reduces class imbalance, and analyzes the results in comparison with participants in the BraTS 2024 competition. Results and discussion . The developed model was trained and tested on two datasets of postoperative images of glioblastoma and meningioma. Several metrics are analyzed to compare the model with the methods described in the literature, as well as to evaluate it in the context of the variability of manual segmentation by different experts. The model achieves a Sorensen coefficient 0.8299 for meningioma segmentation and 0.7028 for glioblastoma contrast-accumulating region segmentation. In addition, the segmentation of the model provides an accurate estimate of the volume of the tumor area, as evidenced by the high values of the coefficient of intra-class correlation – 0.9661 for meningioma and 0.8339 for glioblastoma. Overall, the developed model requires fewer resources for learning and getting results. Conclusions. Model performs segmentation at least at the expert level, but with significantly less variability, especially when assessing the volume of the tumor after surgical treatment.

Short stature is a common concern that necessitates pediatric endocrinology evaluation. Growth hormone (GH) deficiency is often included as an etiology. Brain and pituitary Magnetic Resonance Imaging (MRI) with gadolinium-based contrast agents (GBCAs) is the imaging modality of choice in assessing patients with GH deficiency. Given the significant strides made in MRI technology that allow improved spatial and contrast resolution, the necessity of using contrast material when obtaining brain and pituitary MRI in cases of short stature and isolated GH deficiency should be reassessed. We preformed a retrospective review of otherwise healthy patients with short stature and/or GH deficiency who underwent brain and pituitary MRI without and with contrast, to assess the benefit of contrast administration. Introduction: Short stature is a common concern that necessitates pediatric endocrinology evaluation. The etiologies of short stature are diverse. GH deficiency is often included as an etiology although it accounts for only 1-2% of short stature cases. The prevalence of GH deficiency is reported to be ~ 1:3500. The vast majority of GH deficiency cases are idiopathic in nature with only 20% due to organic causes. The organic causes of GH deficiency include congenital central nervous system (CNS) anomalies, tumors and other pathologic conditions that involve the pituitary-hypothalamic region. As a result, the radiological assessment of the hypothalamic-pituitary region is considered standard of care for evaluating patients with GH deficiency. Although brain and pituitary MRI is the imaging modality of choice in assessing patients with GH deficiency, its yield in cases of isolated GH deficiency is very low. In a study of 40 otherwise normal patients with isolated GH deficiency, 35 (87.5%) had normal brain MRIs. The abnormal findings of brain MRI in the minority of isolated GH deficiency cases included pituitary hypoplasia, pituitary stalk agenesis, lack of the normal T1-weighted pituitary hyperintensity in the posterior part of the sella turcica, and the presence of a high-intensity signal at the infundibular level representing ectopic neurohypophysis. Traditionally, these brain and pituitary MRI images are obtained with the use of contrast material (gadolinium). The main purpose of using contrast material is for the evaluation of pituitary microadenomas. Given the fact that significant strides made in MRI technology and pituitary microadenomas are not appeared to be associated with GH deficiency, the necessity of using contrast material when obtaining brain and pituitary MRI in cases of short stature and isolated GH deficiency should be reassessed. GBCAs have been shown to deposit in different tissues including the kidneys and the brain. The risk increases with repeated doses. The clinical significance of this deposition is unclear at this time but warrants caution especially in pediatric population who have a longer expected lifespan to manifest any delayed effects. Allergic reactions and gastrointestinal symptoms in pediatric patients can occur with GBCA administration, although the incidence is low. Using contrast material also increases the total cost of the MRI study and prolongs the time needed to complete it. Moreover, in order to use contrast material, intravenous venous (IV) access is required which causes discomfort and additional stress to children and their families. Therefore, we performed a retrospective review of otherwise healthy patients with short stature and/or growth hormone deficiency who underwent brain and pituitary MRI without and with contrast, to assess whether contrast administration led to diagnoses that would have otherwise been missed and/or impacted the patient’s clinical course.Objectives: - To compare the diagnostic yield of non-contrast MRI with pre and post-contrast MRI of the brain and pituitary in evaluation of pediatric patients with short stature and/or growth hormone deficiency.- A secondary objective is to measure the size of the pituitary gland and correlates it with peak growth hormone levels (using insulin/argenine). Methodology: We included patients who underwent brain/pituitary MRI with/without contrast performed at our institution between Jan 2013-Dec 2018 who have short stature/GH deficiency. We excluded patients with known diagnosis of other pituitary hormone deficiencies prior to obtaining MRI studies, genetic and neurological disorders, known tumors/malignancies of any type, or renal failure. Two pediatric neuroradiologists independently reviewed the brain and pituitary MRI of these patients (each read 50% of the cohort) blinded to the clinical data and diagnoses. Each radiologist initially reviewed only the non-contrast portions of the studies, and subsequently, the same radiologist reviewed the entire study, including pre- and post-contrast portions in a separate session. The two sessions were 6 weeks apart to avoid recall bias. Several imaging findings including size and morphology of pituitary gland, presence of congenital anomalies or focal lesions and any associated intracranial findings systematically recorded, and subsequently analyzed. Hypotheses: 1.The incidence of finding congenital pituitary cysts is the same when obtaining brain/pituitary MRI imaging using gadolinium contrast versus when not using contrast in patients with short stature and or isolate GH deficiency. 2.The incidence of discovering abnormal infundibulum is the same when obtaining brain/pituitary MRI imaging using gadolinium contrast versus when not using contrast in patients with short stature and or isolate GH deficiency. 3.Small pituitary size correlate with GH deficiency. Results: -We identified 327 patients with short stature/GH deficiency from Jan 2013-Dec 2018-224 (68.5%) are males and 103 (31.5%) are females. -The mean age at the time of imaging is 10 years and the median is 11 years. -161 (49.24%) have height z-score &amp;lt; -2.25 and 166 (50.76%) have height z-score &amp;gt; -2.25.-82 (25.07%) have IGF1 z-score for age &amp;lt; -2, 102 (31.19%) have z-score ≥-2 to ≤ -1, 141 (43.12%) have z-score &amp;gt; -1 and 2 (0.62%) have no level done.-63 (19.27%) have GH peak &amp;lt;5, 87 (26.61%) have GH peak 5-7.99, 53 (16.21%) have GH peak 8-9.99, 30 (9.17%) have GH peak &amp;gt; 10 and 94 (28.75%) did not undergo GH provocative testing. -The kappa coefficient for pars intermedia cyst on pre vs. post contrast imaging is 0.74 and 0.55 for the infundibulum on pre vs. post contrast imaging. -The mean pituitary height for patients with IGF z-score &amp;lt; -2 is 3.9 mm, 4 mm for z-score ≥- 2 to ≤ -1 and 4.3 mm for z-score &amp;gt; -1-The mean pituitary height for patients with peak GH &amp;lt; 5 is 3.8 mm, 4.2 mm for peak 5-7.99, 4.3 mm for peak 8-9.99 and 4.4 mm for peak &amp;gt; 10. Conclusion: This question has not been answered or even raised in the literature. Our findings suggest that the there is no added benefit to use gadolinium when obtaining brain/pituitary MRI for the evaluation of GH deficiency/short stature. Furthermore, it seems that there is an association between the pituitary height and the GH status of the cohort which is in line with previous published studies.

With the success of convolutional neural network (CNN) in the field of image processing and image recognition, an improved fully convolutional neural network (FCNN) based on CNN is proposed. Although FCNN has better performance than CNN, the number of parameters of FCNN has also increased significantly compared to CNN. Consequently, this paper proposes a multi‐feature fusion decision convolutional neural network (MFFD‐CNN), which use two kinds down‐sample methods in the pooling layer, and chooses the concatenate or add operation as the way that consolidates the feature of the convolutional layer with a stride of 2 and the max‐pooling layer to execute fusion operations. Meanwhile replaces the traditional softmax loss with an additive margin Softmax (AM‐Softmax) loss. The structure, which is based on the multi‐target classification experiment results of the MSTAR data set, shows can not only obtain an average correct recognition rate 1.2% higher than FCNN, but also better stability without augmenting the training samples of MSTAR data.

This research study is to detect the Epileptic seizures using Novel Multi-Layered Convolutional Neural Networks (NMLCNN) and to compare the performance with Fully Convolutional Neural Networks (FCNN). By using Novel Multi-Layered Convolution Neural Networks algorithm works with the data collected from the observations of nervous system cells. The changes in the Epileptic seizures from the resource’s constraint, when the criteria met with randomizing, this will arrange in random ways for n times. The most typical beginning of a seizure is detectable through analysis of the Electroencephalogram (EEG). To examine this, the study utilized 20 samples in each group, with a Gpower value of 80%. The study used a predefined significance value and a Confidence Interval (CI) of 95%.This experiment found that NMLCNN has 92.25% and FCNN has 87.55% of accuracy for epileptic seizure. The Novel Multi-Layered Convolutional Neural Network achieved 0.012 significance using group analysis and provided significance. This probing concludes based on accuracy and significance results, detection of Epileptic seizures using Novel Multi-layered Convolutional Neural Networks has performed better compared to Fully Convolutional Neural Network.

In this paper, a new method is proposed to decrease sensitiveness to motion noise and uncertainty in magnetic resonance imaging (MRI) segmentation especially when only one brain image is available. The method is approached with considering spatial neighborhood information by fusing the information of pixels with their neighbors with Dempster–Shafer (DS) theory. The basic probability assignment (BPA) of each single hypothesis is obtained from the membership function of applying fuzzy c-means (FCM) clustering to the gray levels of the MRI. Then multiple hypotheses are generated according to the single hypothesis. Then we update the objective pixel’s BPA by fusing the BPA of the objective pixel and those of its neighbors to get the final result. Some examples in MRI segmentation are demonstrated at the end of the paper, in which our method is compared with some previous methods. The results show that the proposed method is more effective than other methods in motion-blurred MRI segmentation.

Organ-at-risk (OAR) segmentation is a key step for radiotherapy treatment planning. Model-based segmentation (MBS) has been successfully used for the fully automatic segmentation of anatomical structures and it has proven to be robust to noise due to its incorporated shape prior knowledge. In this work, we investigate the advantages of combining neural networks with the prior anatomical shape knowledge of the model-based segmentation of organs-at-risk for brain radiotherapy (RT) on Magnetic Resonance Imaging (MRI). We train our boundary detectors using two different approaches: classic strong gradients as described in [4] and as a locally adaptive regression task, where for each triangle a convolutional neural network (CNN) was trained to estimate the distances between the mesh triangles and organ boundary, which were then combined into a single network, as described by [1]. We evaluate both methods using a 5-fold cross- validation on both T1w and T2w brain MRI data from sixteen primary and metastatic brain cancer patients (some post-surgical). Using CNN-based boundary detectors improved the results for all structures in both T1w and T2w data. The improvements were statistically significant (p < 0.05) for all segmented structures in the T1w images and only for the auditory system in the T2w images.

With the leaps and bounds of computer performance and the advent of the era of big data, deep learning has drawn more and more attention from all walks of life. It can combine low-level features to form more abstract high-level features and describe the data more essentially. Therefore, it is widely used in various fields such as computer vision. Image segmentation is one of the most basic research topics in the field of computer vision. The main purpose is to extract regions of interest from images for later image processing research. In 3D reconstruction based on sequence images, the segmentation accuracy and speed of sequence images determine the quality and efficiency of target reconstruction. Therefore, when facing large-scale sequence images, the biggest problem is how to improve the segmentation speed while ensuring accuracy. Based on the above background, the research content of this article is an indoor scene segmentation algorithm based on full convolutional neural network. According to the characteristics of indoor application scenes, this paper proposes a fast convolutional neural network image segmentation method to segment the indoor scene image and construct the fast fully convolutional networks (FFCN) for indoor scene image segmentation uses inter-layer fusion to reduce the amount of network calculation parameters and avoid the loss of picture feature information by continuous convolution. In order to verify the effectiveness of the network, in this paper, a basic living object data set (XAUT data set) in an indoor environment is created. The XAUT data set is used to train the FFCN network under the Caffe framework to obtain an indoor scene segmentation model. In order to compare the effectiveness of the model, the structure of the worn FCN8s, FCN16s, and FCN32s models was fine-tuned, and the corresponding algorithm model for indoor scene segmentation was obtained by training with the XAUT data set. The experimental results show that the pixel recognition accuracy of all types of networks has reached 86%, and the mean IU ratio has reached more than 63%. The mean IU of the FCN8s network is the highest at 70.38%, but its segmentation speed is only 1/5 of FFCN. On the premise that other types of indicators are not much different, the average segmentation speed on FFCN fast segmentation convolutional neural network reaches 40 fps. It can be seen that the scale fusion technology can well avoid the loss of image feature information in the network convolution and reddening process. Compared with other FCN networks, it has a faster speed and is conducive to real-time image preprocessing.