Biomedical Engineering: Applications, Basis and Communications

The evolving field of computational image analysis has its applications in the industry, manufacturing and biological sciences, especially in the field of medical imaging. Medical imaging and computational physics have evolved together during the past decades with the advancement in the field of artificial intelligence (AI). Deep learning is the sub-domain of AI that mostly deals with imaging data for classification, segmentation and reconstruction. The time series of medical images of different patients, with different staging are categorized based on the physical and biological consequences. The hypothesis of the current research is that the deep learning tool, if trained on several patients, can identify the stage of cancer swiftly for fresh data sets. During this research, an advance Convolutional Neural Network (CNN) strategy is adopted to classify the cancer stage for a group of patients of gastric cancer. The CNN model makes use of skipping connections for better prediction. CNNs have been quite popular in medical imaging for their ability of feature detection. CNNs are used in the recent literature for the analysis of images. During this research, we have used the state-of-the-art Matlab ResNet CNN toolbox for the analysis of the images obtained from esophageal and gastric cancer patients. It was concluded that RESNET50 is a reliable algorithm for the determination of tumor mass on CT Scans. Moreover, the performance of the model can be improved by giving a comparatively larger data set as an input to the model. Inspired from Caltech101, a logic related to RESNET50 was adopted. The data was processed and an algorithm was designed to develop a mapping, based on the mass of tumor. The algorithm designed successfully identified the images, randomly picked from different patients, based on the image features.

Successful treatment of a patient depends on the accurate determination of the disease type. The advent of big data facilitates the retrieving of medical images and helps physicians in reliable diagnoses using content-based medical image retrieval systems (CBMIR). They consist of a feature extraction module and a distance metric. The extracted textural or deep-based features identify different types of diseases. In the proposed retrieval algorithm, we use the gray level cooccurrence matrix as the common textural characteristics and integrate them with semantic attributes. The semantic features are the geometric characteristics of the tumor that a radiologist employ to distinguish between benign and malignant tumors. These high-level attributes include the Euler number, margin smoothness, and the aspect ratio of the lesion’s size. We used the Minkowski distance measure for computing the similarity of images and applied the proposed algorithm to 200 CT-scan data containing lung lesions obtained from the LIDC database. The types of lesions were benign and malignant. Employing an ablation study, we proved the effectiveness of the semantic feature. The precision of the retrieval results is 93% which is promising compared to recent studies. In the future, we plan to define other kinds of semantic attributes to distinguish stages 1–5 of lung tumors as well.

Manual analysis of retinal images is a complicated and time-consuming task for ophthalmologists. Retinal images are susceptible to non-uniform illumination, poor contrast, transmission error, and noise problems. For the detection of retinal abnormalities, an efficient technique is required that can identify the presence of retinal complications. This paper proposes a methodology to enhance retinal images that use morphological operations to improve the contrast and bring out the fine details in the suspicious region. The enhancement plays a vital role in detecting abnormalities in the retinal images. Luminance gain metric ($G_x)$ is obtained from Gamma correction on luminous channel of $H$*$S$*$V$ (hue, saturation, and value) color model of retinal image to improve luminosity. The efficiency and strength of the proposed methodology are evaluated using the performance evaluation parameters peak signal to noise ratio (PSNR), mean square error (MSE), mean absolute error (MAE), feature structural similarity index metric (FSIM), structural similarity index metric (SSIM), spectral residual index metric (SRSIM), Reyligh feature similarity index metric (RFSIM), absolute mean brightness error (AMBE), root mean square error (RMSE), image quality index (IQI), and visual similarity index (VSI). It has been revealed from the results and statistical analysis using the Friedman test that the proposed method outperforms existing state-of-the-art enhancement techniques.

Age-related Macular Degeneration (AMD) is the prime reason for vision impairment observed in major countries worldwide. Hence an accurate early detection of the disease is vital for more research in this area. Also, having a thorough eye diagnosis to detect AMD is a complex job. This paper introduces a Directed Acyclic Graph (DAG) structure-based Convolutional Neural network (CNN) architecture to better classify Dry or Wet AMD. The DAG architecture can combine features from multiple layers to provide better results. The DAG model also has the capacity to learn multi-level visual properties to increase classification accuracy. Fine tuning of DAG-based CNN model helps in improving the performance of the network. The training and testing of the proposed model are carried out with the Mendeley data set and achieved an accuracy of 99.2% with an AUC value of 0.9999. The proposed model also obtains better results for other parameters such as precision, recall and F1-score. Performance of the proposed network is also compared to that of the related works performed on the same data set. This shows ability of the proposed method to grade AMD images to help early detection of the disease. The model also performs computationally efficient for real-time applications as it does the classification process with few learnable parameters and fewer Floating-Point Operations (FLOPs).

The intervertebral disc (IVD) is the soft tissue between the vertebral bodies, which is responsible for transmitting multi-directional loads through the spine and to allow relative motion between the vertebral bodies. The IVD is composed of three distinct tissues, including the nucleus pulposus, annulus fibrosus, and the cartilaginous endplates. Each of these tissues has a characteristic composition and structure which provide them with unique mechanical properties. Among these, nucleus pulposus and annulus fibrosus due to their intricate time-dependent mechanical response has always been the topic of interest for the researchers. Here, we aimed at establishing a patient-specific 3D finite element (FE) model of human IVD based on the poroviscoelastic constitutive law. The main objective was to use the data of tensile stress-relaxation tests on the annulus and nucleus regions to find the poroviscoelastic material constitutive law. The model assumed that the disc is a two-phase body consisting of a water-saturated solid matrix. To do that, the available data in the literature was used as the primary material properties of our model. Thereafter, a set of compressive and tensile loadings was applied on the established patient-specific model of the IVD and the FE results of the poroviscoelastic model were compared to the experimental data. This allowed us to determine a new set of revised parameter values for the poroviscoelastic model which will have practical implications for any future FE studies.

This paper reports the innovative results of applying nonlinear dynamical theory to heart-rate (HR) and respiratory (RP) sequences to explore Yin–Yang (passive–active) interacting phenomenon of cardiorespiratory function. Organ systems inside human body interact mutually in the manner of swapping the active–passive role, according to Yin–Yang theory in traditional Chinese medicine. Yang as an active source rules over Yin as a passive load. Nonlinear inter-driven (NLID) factor, ${\mathrm{\Gamma }}_{(Y/X)}$, based on recurrence analysis is proposed in this study. NLID factor characterizes the nonlinear interactions between two nonlinear dynamical systems, cardiac and respiratory systems. NLID factor is computed by quantifying the HR and RP recurrent plots constructed from phase trajectories based on embedding theorem and phase-space reconstruction. Heterogeneous patterns of HR and RP recurrence plots in continuous-attention task reflect highly unpredictable system dynamics. Heterogeneous HR recurrence patterns are surprisingly associated with the nearly homogeneous RP recurrence plot at resting state, indicating rather weak cardiorespiratory interaction. HR recurrence in Zen exhibits notable regularity in spite of the irregular RP recurrence behaviors. NLID factor reveals prominent Yang energy delivered by RP system that dominates over the cardiac system in Zen-meditation group. Such findings draw our attention to the possibly extraordinary self-regularity power of heart (and autonomic nervous system) in the Zen-meditation state.

Nowadays, the number of patients with brain tumors is steadily increasing, diagnosis and isolation of the tumor play an important role in the process of treatment and surgery. Due to the high error of manual segmentation of the tumor, algorithms that perform this operation with less error are of great importance. Convolutional neural networks have made great progress in the field of medical imaging. The use of imaging techniques and pattern recognition in the diagnosis and automatic determination of brain tumors by MRI imaging reduces errors, human error and speeds up detection. The artificial convolutional neural network (CNN) has been widely used in the diagnosis of intelligent cancers and has significantly reduced the error rate. Therefore, in this paper, we present a new method using a combination of convolutional and multi-scale artificial neural network that has significantly increased the accuracy of tumor diagnosis. This study presents a multidisciplinary convolution neural network (MCNN) approach to classifying tumors that can be used as an important part of automated diagnosis systems for accurate cancer diagnosis. Based on the MCNN structure, which presents the MRI image to several deep convolutional neural networks of varying sizes and resolutions, the stage of extracting classical hand-made features is avoided. This approach proposes better classification rates than the classical methods. This study uses a multi-scale convolution technique to achieve a detection accuracy of 95/4%, which shows the efficiency of the proposed method.

Flying robots popularly known as drones or UAVs are emerging technologies of the current era. A significant amount of research work has been undertaken in this area in the last few years. Considering the current scenario where aerial vehicles are taking a major part of the market it is important to have an effective and robust design of flying robots. This paper aims to examine the categories of flying robots based on the features that include a range from petite to large and its body structure, wing designs, tail design, propulsion system, and gripper mechanisms along with the associated materials and manufacturing techniques. Again the work is intended to review the respective challenges faced by each category. Mostly the challenges faced by flying robots are design challenges, material selection, and fabrication challenges which are discussed in the paper. In this paper, we have summarized various designs of flying robots developed to date as well as we have focused on major features to be taken care of while designing flying robots. This paper has tried to focus on different design aspects and challenges faced by flying robots so that further research can be carried out to develop effective flying robots in the future.