Date of Graduation


Document Type


Degree Name

Bachelor of Science

Degree Level



Computer Science and Computer Engineering


Li, Qinghua

Committee Member/Reader

Patitz, Matthew

Committee Member/Second Reader

Luu, Khoa


The proliferation of modern smartphone camera use in the past decade has resulted in unprecedented numbers of personal photos being taken and stored on popular devices. However, it has also caused privacy concerns. These photos sometimes contain potentially harmful information if they were to be leaked such as the personally identifiable information found on ID cards or in legal documents. With current security measures on iOS and Android phones, it is possible for 3rd party apps downloaded from official app stores or other locations to access the photo libraries on these devices without user knowledge or consent. Additionally, the prevalence of smartphone cameras in public has reduced personal privacy, as strangers are commonly photographed without permission. To mitigate the privacy risk posed by apps and unwanted public photos, this research project explores 3 main topics: developing a two-step method including permission analysis and system call analysis to identify the possibility of 3rd party applications accessing sensitive photos without user knowledge, developing an automated classifier to identify and protect private photos in smartphone media storage, and creating an accurate computer vision model for identifying bystanders in photos, so that their faces might be later blurred or otherwise obfuscated to protect their privacy. The resulting data from the system call analysis will hopefully improve public awareness on the vulnerabilities created by downloading untrustworthy apps. The private photo classifier and bystander detection model are able to achieve acceptable accuracy on the test datasets and can be used in future works to implement working systems to protect individual privacy in the aforementioned threat cases.


Privacy, Smartphones, Android, iOS, Photos, Images