Catherine Todd

Force Application During Cochlear Implant Insertion: An Analysis for Improvement of Surgeon Technique

Highly invasive surgical procedures, such as the implantation of a prosthetic device, require correct force delivery to achieve desirable outcomes and minimize trauma induced during the operation. Improvement in surgeon technique can reduce the chances of excessive force application and lead to optimal placement of the electrode array. The fundamental factors that affect the degree of success for cochlear implant recipients are identified through empirical methods. Insertion studies are performed to assess force administration and electrode trajectories during implantations of the Nucleusreg 24 Contourtrade and Nucleusreg 24 Contour Advancetrade electrodes into a synthetic model of the human Scala Tympani, using associated methods. Results confirm that the advance off- stylet insertion of the soft-tipped contour advance electrode gives an overall reduction in insertion force. Analysis of force delivery and electrode positioning during cochlear implantation can help identify and control key factors for improvement of insertion method. Based on the findings, suggestions are made to enhance surgeon technique.

Classification Cervical Cancer Using Histology Images

This papers reports on methodologies and outcome of a study aiming at developing robust tool to evaluate and classify histology images of cervical cancer. Using the histology images acquired from the pathology laboratories in an Indonesian hospital, this study aims to classify cervical biopsy images based on four well known discriminatory features a) the ratio of nuclei to cytoplasm b) diameter of nuclei c) shape factor and d) roundness of nuclei. In this study, the cervical histology images are classified into three categories: 1) normal, 2) pre cancer and 3) malignant. The final system will take as input a biopsy image of the image of the cervix containing the epithelium layer and provide the classification using the new automated approach, to assist the pathologist in cervical cancer diagnosis.

Cervical Cancer Classification Using Gabor Filters

This paper presents a novel algorithm for computer-assisted classification of cervical cancers using digitized histology images of biopsies. Texture analysis of the nuclei structure is very important for classification of cervical cancer histology. In this paper we present a two-tier classification strategy using Gabor filter banks for local classification and abnormality spread for global taxonomy. The test data used in this work are digitized histology images of cervical biopsies acquired from the pathology laboratories in the Saiful An war Hospital in Indonesia. The images from over 500 subjects are categorized by the pathologists into five grades, benign, pre-cancer (CIN1, CIN2, CIN3) and malignant. In the algorithm developed in this work, a texture classification method using Gabor filter banks is implemented to segment the image into five possible regions: of background, normal, abnormal, basal and stroma cells. The global classification algorithm uses the segmented image for the final prognosis of the degree of malignancies from benign to malignant. The process of texture segmentation using the Gabor filter bank involves the application of filters for several spatial frequencies and orientations. The Gabor filter bank is applied to cervical histology images with six frequencies and four orientations. Feature vectors are formed, comprising the response of each pixel and its neighboring pixels to each filter. The feature vectors are then used to classify each pixel and its immediate neighbor pixels into five categories. Based on the spread of abnormalities on the epithelium layer, the cervical histology image is then classified. The classification results are then used to further classify the image into: 1) normal, 2) pre-cancer and 3) malignant. The pre-cancer is divided into: a) CIN 1, b) CIN 2 and c) CIN 3. The final system will take as input a biopsy image of the cervix containing the epithelium layer and provide the classification using our new approach, to assist the pathologist in cervical cancer diagnosis.

Virtual cochlear implant insertion for medical education

A surgical simulator has been developed for the purpose of training otologists in cochlear implantation. The simulation provides real-time visual and haptic feedback during implant insertion into the human Scala Tympani (ST). The benefits and possible outcomes for this type of simulator are presented. Methods for model generation are discussed, for anatomical and prosthetic structures used in the simulation. Development of the interactive model with force-feedback is presented, with results.

Reinforcement learning neural network (RLNN) based adaptive control of fine hand motion rehabilitation robot

Recent neural science research suggests that a robotic device can be an effective tool to deliver the repetitive movement training that is needed to trigger neuroplasticity in the brain following neurologic injuries such as stroke and spinal cord injury (SCI). In such scenario, adaptive control of the robotic device to provide assistance as needed along the intended motion trajectory with exact amount of force intensity, though complex, is a more effective approach. A critic-actor based reinforcement learning neural network (RLNN) control method is explored to provide adaptive control during post-stroke fine hand motion rehabilitation training. The effectiveness of the method is verified through computer simulation and implementation on a hand rehabilitation robotic device. Results suggest that the control system can fulfil the assist-as-needed (AAN) control with high performance and reliability. The method demonstrates potential to encourage active participation of the patient in the rehabilitation process and to improve the efficiency of the process.

Unsupervised Image Classification by Probabilistic Latent Semantic Analysis for the Annotation of Images

Image annotation has been identified to be a suitable means by which the semantic gap which has made the accuracy of Content-based image retrieval unsatisfactory be eliminated. However existing methods of automatic annotation of images depends on supervised learning, which can be difficult to implement due to the need for manually annotated training samples which are not always readily available. This paper argues that the unsupervised learning via Probabilistic Latent Semantic Analysis provides a more suitable machine learning approach for image annotation especially due to its potential to based categorisation on the latent semantic content of the image samples, which can bridge the semantic gap present in Content Based Image Retrieval. This paper therefore proposes an unsupervised image categorisation model in which the semantic content of images are discovered using Probabilistic Latent Semantic Analysis, after which they are clustered into unique groups based on semantic content similarities using K-means algorithm, thereby providing suitable annotation exemplars. A common problem with categorisation algorithms based on Bag-of-Visual Words modelling is the loss of accuracy due to spatial incoherency of the Bag-of-Visual Word modelling, this paper also examines the effectiveness of Spatial pyramid as a means of eliminating spatial incoherency in Probabilistic Latent Semantic Analysis classification.

Morphological characteristics of cervical cells for cervical cancer diagnosis

This paper investigates cervical cancer diagnosis based on the morphological characteristics of cervical cells. The developed algorithms cover several steps: pre-processing, image segmentation, nuclei and cytoplasm detection, feature calculation, and classification. The K-means clustering algorithm based on colour segmentation is used to segment cervical biopsy images into five regions: background, nuclei, red blood cell, stroma and cytoplasm. The morphological characteristics of cervical cells are used for feature extraction of cervical histopathology images. The cervical histopathology images are classified using four well known discriminatory features: 1) the ratio of nuclei to cytoplasm, 2) the diameter of nuclei, 3) the shape factor and 4) the compactness of nuclei. Finally, the images are analysed and classified into appropriate classes. This method is utilised to classify the cervical biopsy images into normal, pre-cancer (Cervical Intraepithelial Neoplasia (CIN)1, CIN2, CIN3) and malignant.

Visual and Haptic Rendering of Temporal Bone Surgery

Abstract implantation. Visual and haptic rendering of temporal bone drilling comprises the second stage of system design and is the focus of this paper. A geometric model has been produced for this purpose. The suitability of the Reachin Application Programming Interface (API) to facilitate simulator implementation is examined. Texture mapping is applied for visual rendering of drill bit and bone interactions. Force and torque feedback associated with bone dissection are introduced. The progress made is reported. The model will incorporate electrode array insertion as a final stage in design.

The Combined Effects of Adaptive Control and Virtual Reality on Robot-Assisted Fine Hand Motion Rehabilitation in Chronic Stroke Patients: A Case Study

Robot-assisted therapy is regarded as an effective and reliable method for the delivery of highly repetitive training that is needed to trigger neuroplasticity following a stroke. However, the lack of fully adaptive assist-as-needed control of the robotic devices and an inadequate immersive virtual environment that can promote active participation during training are obstacles hindering the achievement of better training results with fewer training sessions required. This study thus focuses on these research gaps by combining these 2 key components into a rehabilitation system, with special attention on the rehabilitation of fine hand motion skills. The effectiveness of the proposed system is tested by conducting clinical trials on a chronic stroke patient and verified through clinical evaluation methods by measuring the key kinematic features such as active range of motion (ROM), finger strength, and velocity. By comparing the pretraining and post-training results, the study demonstrates that the proposed method can further enhance the effectiveness of fine hand motion rehabilitation training by improving finger ROM, strength, and coordination.

Efficient region of interest detection using blind image division

The determination of Region-of-Interest can be used as a means of improving the performance of image retrieval, when used in image annotation as a step in the indexing of images collection. It also has the potential to support efficient video compression for real-time applications. However, existing Region-of-Interest detection methods are mostly unsuitable for managing large number of images and for real-time video applications due to their high computational requirements. This paper therefore proposes an unsupervised algorithm which applies blind image division in the determination of relevant regions within an image space.