Deep Learning Paradigm and Its Bias for Coronary Artery Wall Segmentation in Intravascular Ultrasound Scans: A Closer Look

BACKGROUND AND MOTIVATION

Coronary artery disease (CAD) has the highest mortality rate; therefore, its diagnosis is vital. Intravascular ultrasound (IVUS) is a high-resolution imaging solution that can image coronary arteries, but the diagnosis software via wall segmentation and quantification has been evolving. In this study, a deep learning (DL) paradigm was explored along with its bias.

METHODS

Using a PRISMA model, 145 best UNet-based and non-UNet-based methods for wall segmentation were selected and analyzed for their characteristics and scientific and clinical validation. This study computed the coronary wall thickness by estimating the inner and outer borders of the coronary artery IVUS cross-sectional scans. Further, the review explored the bias in the DL system for the first time when it comes to wall segmentation in IVUS scans. Three bias methods, namely (i) ranking, (ii) radial, and (iii) regional area, were applied and compared using a Venn diagram. Finally, the study presented explainable AI (XAI) paradigms in the DL framework.

FINDINGS AND CONCLUSIONS

UNet provides a powerful paradigm for the segmentation of coronary walls in IVUS scans due to its ability to extract automated features at different scales in encoders, reconstruct the segmented image using decoders, and embed the variants in skip connections. Most of the research was hampered by a lack of motivation for XAI and pruned AI (PAI) models. None of the UNet models met the criteria for bias-free design. For clinical assessment and settings, it is necessary to move from a paper-to-practice approach.

Ensemble Deep Learning Derived from Transfer Learning for Classification of COVID-19 Patients on Hybrid Deep-Learning-Based Lung Segmentation: A Data Augmentation and Balancing Framework

BACKGROUND AND MOTIVATION

Lung computed tomography (CT) techniques are high-resolution and are well adopted in the intensive care unit (ICU) for COVID-19 disease control classification. Most artificial intelligence (AI) systems do not undergo generalization and are typically overfitted. Such trained AI systems are not practical for clinical settings and therefore do not give accurate results when executed on unseen data sets. We hypothesize that ensemble deep learning (EDL) is superior to deep transfer learning (TL) in both non-augmented and augmented frameworks.

METHODOLOGY

The system consists of a cascade of quality control, ResNet-UNet-based hybrid deep learning for lung segmentation, and seven models using TL-based classification followed by five types of EDL's. To prove our hypothesis, five different kinds of data combinations (DC) were designed using a combination of two multicenter cohorts-Croatia (80 COVID) and Italy (72 COVID and 30 controls)-leading to 12,000 CT slices. As part of generalization, the system was tested on unseen data and statistically tested for reliability/stability.

RESULTS

Using the K5 (80:20) cross-validation protocol on the balanced and augmented dataset, the five DC datasets improved TL mean accuracy by 3.32%, 6.56%, 12.96%, 47.1%, and 2.78%, respectively. The five EDL systems showed improvements in accuracy of 2.12%, 5.78%, 6.72%, 32.05%, and 2.40%, thus validating our hypothesis. All statistical tests proved positive for reliability and stability.

CONCLUSION

EDL showed superior performance to TL systems for both (a) unbalanced and unaugmented and (b) balanced and augmented datasets for both (i) seen and (ii) unseen paradigms, validating both our hypotheses.

Eight pruning deep learning models for low storage and high-speed COVID-19 computed tomography lung segmentation and heatmap-based lesion localization: A multicenter study using COVLIAS 2.0

BACKGROUND

COVLIAS 1.0: an automated lung segmentation was designed for COVID-19 diagnosis. It has issues related to storage space and speed. This study shows that COVLIAS 2.0 uses pruned AI (PAI) networks for improving both storage and speed, wiliest high performance on lung segmentation and lesion localization.

METHOD

ology: The proposed study uses multicenter ∼9,000 CT slices from two different nations, namely, CroMed from Croatia (80 patients, experimental data), and NovMed from Italy (72 patients, validation data). We hypothesize that by using pruning and evolutionary optimization algorithms, the size of the AI models can be reduced significantly, ensuring optimal performance. Eight different pruning techniques (i) differential evolution (DE), (ii) genetic algorithm (GA), (iii) particle swarm optimization algorithm (PSO), and (iv) whale optimization algorithm (WO) in two deep learning frameworks (i) Fully connected network (FCN) and (ii) SegNet were designed. COVLIAS 2.0 was validated using "Unseen NovMed" and benchmarked against MedSeg. Statistical tests for stability and reliability were also conducted.

RESULTS

Pruning algorithms (i) FCN-DE, (ii) FCN-GA, (iii) FCN-PSO, and (iv) FCN-WO showed improvement in storage by 92.4%, 95.3%, 98.7%, and 99.8% respectively when compared against solo FCN, and (v) SegNet-DE, (vi) SegNet-GA, (vii) SegNet-PSO, and (viii) SegNet-WO showed improvement by 97.1%, 97.9%, 98.8%, and 99.2% respectively when compared against solo SegNet. AUC > 0.94 (p < 0.0001) on CroMed and > 0.86 (p < 0.0001) on NovMed data set for all eight EA model. PAI <0.25 s per image. DenseNet-121-based Grad-CAM heatmaps showed validation on glass ground opacity lesions.

CONCLUSIONS

Eight PAI networks that were successfully validated are five times faster, storage efficient, and could be used in clinical settings.

A novel cloning template designing method by using an artificial bee colony algorithm for edge detection of CNN based imaging sensors

Cellular Neural Networks (CNNs) have been widely used recently in applications such as edge detection, noise reduction and object detection, which are among the main computer imaging processes. They can also be realized as hardware based imaging sensors. The fact that hardware CNN models produce robust and effective results has attracted the attention of researchers using these structures within image sensors. Realization of desired CNN behavior such as edge detection can be achieved by correctly setting a cloning template without changing the structure of the CNN. To achieve different behaviors effectively, designing a cloning template is one of the most important research topics in this field. In this study, the edge detecting process that is used as a preliminary process for segmentation, identification and coding applications is conducted by using CNN structures. In order to design the cloning template of goal-oriented CNN architecture, an Artificial Bee Colony (ABC) algorithm which is inspired from the foraging behavior of honeybees is used and the performance analysis of ABC for this application is examined with multiple runs. The CNN template generated by the ABC algorithm is tested by using artificial and real test images. The results are subjectively and quantitatively compared with well-known classical edge detection methods, and other CNN based edge detector cloning templates available in the imaging literature. The results show that the proposed method is more successful than other methods.

Parallel Genetic-Based Hybrid Pattern Matching Algorithm for Isolated Word Recognition

Dynamic Time Warping (DTW) is a common technique widely used for nonlinear time normalization of different utterances in many speech recognition systems. Two major problems are usually encountered when the DTW is applied for recognizing speech utterances: (i) the normalization factors used in a warping path; and (ii) finding the K-best warping paths. Although DTW is modified to compute multiple warping paths by using the Tree-Trellis Search (TTS) algorithm, the use of actual normalization factor still remains a major problem for the DTW. In this paper, a Parallel Genetic Time Warping (PGTW) is proposed to solve the above said problems. A database extracted from the TIMIT speech database of 95 isolated words is set up for evaluating the performance of the PGTW. In the database, each of the first 15 words had 70 different utterances, and the remaining 80 words had only one utterance. For each of the 15 words, one utterance is arbitrarily selected as the test template for recognition. Distance measure for each test template to the utterances of the same word and to those of the 80 words is calculated with three different time warping algorithms: TTS, PGTW and Sequential Genetic Time Warping (SGTW). A Normal Distribution Model based on Rabiner23 is used to evaluate the performance of the three algorithms analytically. The analyzed results showed that the PGTW had performed better than the TTS. It also showed that the PGTW had very similar results as the SGTW, but about 30% CPU time is saved in the single processor system.

SEGMENTATION OF OVARIAN ULTRASOUND IMAGES USING CELLULAR NEURAL NETWORKS

Segmentation of ovarian ultrasound images using cellular neural networks (CNNs) is studied in this paper. The segmentation method consists of five successive steps where the first four uses CNNs. In the first step, only rough position of follicles is determined. In the second step, the results are improved by expansion of detected follicles. In the third step, previously undetected inexpressive follicles are determined, while the fourth step detects the position of ovary. All results are joined in the fifth step. The templates for CNNs were obtained by applying genetic algorithm. The segmentation method has been tested on 50 ovarian ultrasound images. The recognition rate of follicles was around 60% and misidentification rate was around 30%.

HYBRID APPROACH FOR BRAIN TUMOR SEGMENTATION IN MAGNETIC RESONANCE IMAGES USING CELLULAR NEURAL NETWORKS AND OPTIMIZATION TECHNIQUES

Tumor segmentation from brain magnetic resonance image data is an important but time consuming task performed manually by medical experts. Automating this process is challenging due to the high diversity in appearance of tumor tissue among different patients and in many cases, similarity between tumor and normal tissue. This paper deals with an efficient segmentation algorithm for extracting brain tumors in magnetic resonance images using Cellular Neural Networks (CNN). Learning CNN templates values are formulated as an optimization problem. The template coefficients (weights) of an CNN which will give a desired performance, can be derived by learning genetic algorithm and simulated annealing optimization techniques. The objective of this work is to compare the performance of genetic algorithm (GA) and simulated annealing (SA) for finding the optimum template values in the CNN which is used for segmenting the tumor region in the abnormal MR images. The method is applied on real data of MRI images of thirty patients with four different types of tumors. The results are compared with radiologist labeled ground truth. Quantitative analysis between ground truth and segmented tumor is presented in terms of segmentation efficiency. From the analysis and performance measures like segmentation accuracy, it is inferred that the brain tumor segmentation is best done using CNN with genetic algorithm template optimization than CNN with simulated annealing template optimization. An average accuracy rate of above 95% was obtained using this segmentation algorithm.

Nodule detection in a lung region that's segmented with using genetic cellular neural networks and 3D template matching with fuzzy rule based thresholding

Objective

The purpose of this study was to develop a new method for automated lung nodule detection in serial section CT images with using the characteristics of the 3D appearance of the nodules that distinguish themselves from the vessels.

Materials and Methods

Lung nodules were detected in four steps. First, to reduce the number of region of interests (ROIs) and the computation time, the lung regions of the CTs were segmented using Genetic Cellular Neural Networks (G-CNN). Then, for each lung region, ROIs were specified with using the 8 directional search; +1 or -1 values were assigned to each voxel. The 3D ROI image was obtained by combining all the 2-Dimensional (2D) ROI images. A 3D template was created to find the nodule-like structures on the 3D ROI image. Convolution of the 3D ROI image with the proposed template strengthens the shapes that are similar to those of the template and it weakens the other ones. Finally, fuzzy rule based thresholding was applied and the ROI's were found. To test the system's efficiency, we used 16 cases with a total of 425 slices, which were taken from the Lung Image Database Consortium (LIDC) dataset.

Results

The computer aided diagnosis (CAD) system achieved 100% sensitivity with 13.375 FPs per case when the nodule thickness was greater than or equal to 5.625 mm.

Conclusion

Our results indicate that the detection performance of our algorithm is satisfactory, and this may well improve the performance of computer-aided detection of lung nodules.

Estimating nonlinear interdependences in dynamical systems using cellular nonlinear networks

We propose a method for estimating nonlinear interdependences between time series using cellular nonlinear networks. Our approach is based on the nonlinear dynamics of interacting nonlinear elements. We apply it to time series of coupled nonlinear model systems and to electroencephalographic time series from an epilepsy patient, and we show that an accurate approximation of symmetric and asymmetric realizations of a nonlinear interdependence measure can be achieved, thus allowing one to detect the strength and direction of couplings.

Automated Detection of Small-Size Pulmonary Nodules Based on Helical CT Images

A computer-aided diagnosis (CAD) system to detect small-size (from 2 mm to around 10 mm) pulmonary nodules in helical CT scans is developed. This system uses different schemes to locate juxtapleural nodules and non-pleural nodules. For juxtapleural nodules, morphological closing, thresholding and labeling are performed to obtain volumetric nodule candidates; gray level and geometric features are extracted and analyzed using a linear discriminant analysis (LDA) classifier. To locate non-pleural nodules, a discrete-time cellular neural network (DTCNN) uses local shape features which successfully capture the differences between nodules and non-nodules, especially vessels. The DTCNN was trained using genetic algorithm (GA). Testing on 17 cases with 3979 slice images showed the effectiveness of the proposed system, yielding sensitivity of 85.6% with 9.5 FPs/case (0.04 FPs/image). Moreover, the CAD system detected many nodules missed by human visual reading. This showed that the proposed CAD system acted effectively as an assistant for human experts to detect small nodules and provided a "second opinion" to human observers.

Optimal solutions for cellular neural networks by paralleled hardware annealing

An engineering annealing method for optimal solutions of cellular neural networks is presented. Cellular neural networks are very promising in solving many scientific problems in image processing, pattern recognition, and optimization by the use of stored program with predetermined templates. Hardware annealing, which is a paralleled version of mean-field annealing in analog networks, is a highly efficient method of finding optimal solutions of cellular neural networks. It does not require any stochastic procedure and henceforth can be very fast. The generalized energy function of the network is first increased by reducing the voltage gain of each neuron. Then, the hardware annealing searches for the globally minimum energy state by continuously increasing the gain of neurons. The process of global optimization by the proposed annealing can be described by the eigenvalue problems in the time-varying dynamic system. In typical nonoptimization problems, it also provides enough stimulation to frozen neurons caused by ill-conditioned initial states.