The science of solving clinical problems by analyzing images generated in clinical practice is known as medical image analysis. The aim is to extract information in an effective and efficient manner for improved clinical diagnosis. The recent advances in the field of biomedical engineering has made medical image analysis one of the top research and development area. One of the reason for this advancement is the application of machine learning techniques for the analysis of medical images. Deep learning is successfully used as a tool for machine learning, where a neural network is capable of automatically learning features. This is in contrast to those methods where traditionally hand crafted features are used. The selection and calculation of these features is a challenging task. Among deep learning techniques, deep convolutional networks are actively used for the purpose of medical image analysis. This include application areas such as segmentation, abnormality detection, disease classification, computer aided diagnosis and retrieval. In this study, a comprehensive review of the current state-of-the-art in medical image analysis using deep convolutional networks is presented. The challenges and potential of these techniques are also highlighted.

Medical Image Analysis using Convolutional Neural Networks: A Review

A review on brain tumor segmentation of MRI images.

Brain tumor classification using a hybrid deep autoencoder with Bayesian fuzzy clustering-based segmentation approach

The brain cancer treatment process depends on the physician’s experience and knowledge. For this reason, using an automated tumor detection system is extremely important to aid radiologists and physicians to detect brain tumors. The proposed method has three stages, which are preprocessing, the extreme learning machine local receptive fields (ELM-LRF) based tumor classification, and image processing based tumor region extraction. At first, nonlocal means and local smoothing methods were used to remove possible noises. In the second stage, cranial magnetic resonance (MR) images were classified as benign or malignant by using ELM-LRF. In the third stage, the tumors were segmented. The purpose of the study was using only cranial MR images, which have a mass, in order to save the physician’s time. In the experimental studies the classification accuracy of cranial MR images is 97.18%. Evaluated results showed that the proposed method’s performance was better than the other recent studies in the literature. Experimental results also proved that the proposed method is effective and can be used in computer aided brain tumor detection.

https://journals.tubitak.gov.tr/elektrik/issues/elk-18-26-5/elk-26-5-9-1801-8.pdf

Deep learning based brain tumor classification and detection system

Incipient fault diagnosis of rolling bearings based on adaptive variational mode decomposition and Teager energy operator

Modular self-reconfigurable robots present wide and unique solutions for growing demands in the domains of space exploration, automation, consumer products, and so forth. The higher utilization factor and self-healing capabilities are most demanded traits in robotics for real world applications and modular robotics offer better solutions in these perspectives in relation to traditional robotics. The researchers in robotics domain identified various applications and prototyped numerous robotic models while addressing constraints such as homogeneity, reconfigurability, form factor, and power consumption. The diversified nature of various modular robotic solutions proposed for real world applications and utilization of different sensor and actuator interfacing techniques along with physical model optimizations presents implicit challenges to researchers while identifying and visualizing the merits/demerits of various approaches to a solution. This paper attempts to simplify the comparison of various hardware prototypes by providing a brief study on hardware architectures of modular robots capable of self-healing and reconfiguration along with design techniques adopted in modeling robots, interfacing technologies, and so forth over the past 25 years.

/pdf/modular-self-reconfigurable-robotic-systems-a-survey-on-1j0ho526pl.pdf

Modular Self-Reconfigurable Robotic Systems: A Survey on Hardware Architectures

Ischemic stroke lesion segmentation using stacked sparse autoencoder.

Computerized methods are used in medical imaging to image the inner portions of the human body for medical diagnosis. Image segmentation plays an important role in diagnosis, surgical planning, navigation and various medical evaluations. Manual, semi-automatic and automatic methods are existing for segmentation of the region of interest. In this paper, a hybrid approach for brain tumor detection and classification through magnetic resonance images has been proposed. First phase of the proposed approach deals with image preprocessing which includes noise filtering, skull detection, etc. The second phase deals with feature extraction of MR brain images using gray level co-occurrence matrix. Third phase deals with classification of inputs into normal or abnormal using Least Squares Support Vector Machine classifier with Multilayer perceptron kernel. Final phase is the segmentation of the tumor part from the brain using fast bounding box. The experiments were carried out on 100 images consisting of 25 normal and 75 abnormal from a real human brain and synthetic MRI dataset. The classification accuracy on both training and test images was found to be 96.63%.

Hybrid approach for brain tumor detection and classification in magnetic resonance images

Non-invasive methods accomplished by a computer aided diagnosis of knee-joint disorders provide an effective tool. The objective of this study is to analyse vibroarthographic (VAG) signals using non-linear signal processing technique. This study includes different entropy-based feature extraction techniques to attain highly distinguishable features. The authors proposed to use a non-linear method known as complete ensemble empirical mode decomposition with adaptive white noise to decompose the VAG signals into intrinsic mode functions (IMFs). Entropy-based features involving approximate entropy, sample entropy, Shannon entropy, Renyi entropy, Tsallis entropy and permutation entropy (PeEn) are computed from dominant IMFs and reconstructed VAG signals. These extracted features are given as input to the least squares support vector machine as a classifier. The results illustrated that PeEn performed better with respect to other entropies. PeEn gives a classification accuracy of 86.61% and Matthews correlation coefficient of 0.7082. The computational complexity of entropies was also analysed. Results inferred that PeEn has a computational complexity of O
(
 N
) provided a simple, robust and low computational feature extraction technique. Analysis of VAG signals using non-linear preprocessing and entropy-based features can provide highly distinguishable features for accurate detection of knee-joint disorders.

https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/iet-smt.2017.0284

Entropy-based feature extraction and classification of vibroarthographic signal using complete ensemble empirical mode decomposition with adaptive noise

Anita Agrawal

Papers

Modular Self-Reconfigurable Robotic Systems: A Survey on Hardware Architectures

Ischemic stroke lesion segmentation using stacked sparse autoencoder.

Hybrid approach for brain tumor detection and classification in magnetic resonance images

Entropy-based feature extraction and classification of vibroarthographic signal using complete ensemble empirical mode decomposition with adaptive noise

Time-frequency based feature extraction for the analysis of vibroarthographic signals