No abstract

In this paper, a multipath convolutional neural network (MP-CNN) is proposed for rehabilitation exercise recognition using sensor data. It consists of two novel components: a dynamic convolutional neural network (D-CNN) and a state transition probability CNN (S-CNN). In the D-CNN, Gaussian mixture models (GMMs) are exploited to capture the distribution of sensor data for the body movements of the physical rehabilitation exercises. Then, the input signals and the GMMs are screened into different segments. These form multiple paths in the CNN. The S-CNN uses a modified Lempel–Ziv–Welch (LZW) algorithm to extract the transition probabilities of hidden states as discriminate features of different movements. Then, the D-CNN and the S-CNN are combined to build the MP-CNN. To evaluate the rehabilitation exercise, a special evaluation matrix is proposed along with the deep learning classifier to learn the general feature representation for each class of rehabilitation exercise at different levels. Then, for any rehabilitation exercise, it can be classified by the deep learning model and compared to the learned best features. The distance to the best feature is used as the score for the evaluation. We demonstrate our method with our collected dataset and several activity recognition datasets. The classification results are superior when compared to those obtained using other deep learning models, and the evaluation scores are effective for practical applications.

The MRI or CT scan images are primary follow up diagnostic tools when a neurologic exam indicates a possibility of a primary or metastatic brain tumor existence. The tumor tissue mainly appears in brighter colors than the rest of the regions in the brain. Based on this observation, an automated algorithm for brain tumor detection and medical doctors' assistance in facilitated and accelerated diagnosis procedure has been developed and initially tested on images obtained from the patients with diagnosed tumors and healthy subjects.Keywords --Brain tumor, brain cancer, image processing, segmentationThe existence of a variety of types of tumors with different characteristics necessitates an individually well-defined specialized treatment as outlined by a neuro-oncologist. The cause of brain tumors is unknown but their occurrence is associated with an abnormal growth of cells inside the skull which may lead to serious impairments and/or life-threatening conditions because of their character and the limited space of the intracranial cavity. Research shows that people affected by brain tumors die due to their inaccurate detection [1]. Therefore, an estimated 120 different brain tumors of which some can be benign (noncancerous growths) or malignant (cancerous) require accurate diagnosis which is crucial for an early and successful treatment. The MRI is the most commonly used modality for brain tumor growth imaging and location detection. The conventional method for CT and MRI brain images classification and tumor detection is still mostly based on a direct human inspection of those images, although other methods are being proposed [2-3]. The MRI images visual evaluation and examination by radiologists is subjective by its nature and is time consuming and prone to errors or omissions, however due to the complexity of information at this point, it cannot be substituted with a fully automated evaluation. Therefore algorithmic image processing can assist radiologists in brain tumor diagnosis in multi-parametric MR images, especially since brain tumor detection and segmentation needs to take into account large variations in appearance and shape of structures. Therefore, a computer aided method for automated brain tumor detection in MRI images has been developed. This system for tumor detection and segmentation consists of several stages: 1) Brain MR image is acquired and sharpened by enhancing its contrast; 2) By observing the MR images it can be noted that the skull and meninges are almost of pure white color. Also in most source MR images, it is obvious that the tumor, as well as the skull and meninges have much higher gray level than other tissues and stands out in a much brighter gray level. Noting that the skull appears to be in the shape of a symmetric oval, we removed the skull and meninges region by symmetrically erasing it in four directions: vertically and horizontally. The range of the erasure is controlled by two criteria: grey level and geometrical features. The average gray level of the continuous six pixels is ...

No abstract

No abstract