Reza Samavi

M4CVD: Mobile Machine Learning Model for Monitoring Cardiovascular Disease

Abstract In this paper we present M4CVD: Mobile Machine Learning Model for Monitoring Cardiovascular Disease, a system designed specifically for mobile devices that facilitates monitoring of cardiovascular disease (CVD). The system uses wearable sensors to collect observable trends of vital signs contextualized with data from clinical databases. Instead of transferring the raw data directly to the health care professionals, the system performs analysis on the local device by feeding the hybrid of collected data to a support vector machine (SVM) to monitor features extracted from clinical databases and wearable sensors to classify a patient as “continued risk” or “no longer at risk” for CVD. As a work in progress we evaluate a proof-of-concept M4CVD using a synthetic clinical database of 200 patients. The results of our experiment show the system was successful in classifying a patients CVD risk with an accuracy of 90.5%.

Blockchain Enabled Privacy Audit Logs

Privacy audit logs are used to capture the actions of participants in a data sharing environment in order for auditors to check compliance with privacy policies. However, collusion may occur between the auditors and participants to obfuscate actions that should be recorded in the audit logs. In this paper, we propose a Linked Data based method of utilizing blockchain technology to create tamper-proof audit logs that provide proof of log manipulation and non-repudiation. We also provide experimental validation of the scalability of our solution using an existing Linked Data privacy audit log model.

COC: An Ontology for Capturing Semantics of Circle of Care

Abstract In this paper we propose a circle of care (COC) ontology that specifies concepts and relations necessary to capture a patients circle of care and allows one to make inferences about who is in a patients circle of care. The ontology can improve current access control systems in making decisions regarding access events in real time and help identify past cases of illegitimate access through the access log. We validate the ontology by augmenting the Fast Healthcare Interoperability Resources (FHIR) data model with our COC ontology and presenting an example of access logs from FHIR that is extended to answer circle of care queries.

Tamper-Proof Privacy Auditing for Artificial Intelligence Systems

Privacy audit logs are used to capture the actions of participants in a data sharing environment in order for auditors to check compliance with privacy policies. However, collusion may occur between the auditors and participants to obfuscate actions that should be recorded in the audit logs. In this paper, we propose a Linked Data based method of utilizing blockchain technology to create tamper-proof audit logs that provide proof of log manipulation and non-repudiation.

Timestamp-based Integrity Proofs for Linked Data

In this paper, we first investigate the state-of-the-art methods of generating cryptographic hashes that can be used as an integrity proof for RDF datasets. We then propose an efficient method of computing integrity proofs for Linked Data that constructs a sorted Merkle tree for growing RDF datasets based on timestamps (as a key) that are semantically extractable from the RDF dataset. We evaluate our method by comparing it to existing methods and investigating its resistance to common security threats.

Augmented Reality Based Brain Tumor 3D Visualization

Abstract In this paper we present an augmented reality system for mobile devices that facilitates 3D brain tumor visualization in real time. The system uses facial features to track the subject in the scene. The system performs camera calibration based on the face size of the subject, instead of the common approach of using a number of chessboard images to calibrate the camera every time the application is installed on a new device. Camera 3D pose estimation is performed by finding its position and orientation based on a set of 3D points and their corresponding 2D projections. According to the estimated camera pose, a reconstructed brain tumor model is displayed at the same location as the subject’s real anatomy. The results of our experiment show the system was successful in performing the brain tumor augmentation in real time with a reprojection accuracy of 97%.