The cloud computing and storage have revealed a new technology evolution since its high scalability, high stability and high accessibility. To adapt to the trend of big data, many industrial have moved their IT infrastructures and databases to cloud servers provided by third-party companies. Information privacy is one of the most critical concerns faced by the companies which have moved or are considering moving their data to the cloud. Social media companies, such as Instagram and Facebook, must consider carefully if their data containing user information are safe to be saved on the cloud. To resolve this concern, we have developed an intelligent system - AutoObfu, providing an end-to-end solution for both obfuscating images and classifying the obfuscated images.

With the popularization of smart devices, most people are posting their photos on social networks to share their life with family and friends. Instagram, Facebook, Snapchat have become the most popular applications people use to share their photos nowadays. Most of these applications are using cloud storage to keep photos because it’s high accessibility and low cost. However, the safety and privacy of the photos have become a serious issue due to reports of government monitor public activities online, and many photo leakage of iCloud. Images contain many information about the image uploader, so applications can provide users friend suggestions based on their photos, however, this also makes photo sharing dangerous as attackers can easily get users’ address, personal information, even their interpersonal relationships. Obfuscation of images before uploading them on the cloud becomes necessary to protect user privacy.

• PyQt5 framework for the GUI
• Tensorflow
• Keras framework