SCSE memory

• Main Author: Chong, Shannon Kang Yao
• Other Authors: Tan Ah Hwee
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/76959
• Institution: Nanyang Technological University

The constant evolution of technology plays a huge role in shaping the current society. People are more dependent on technology now than compared to thirty years ago. With these developed technologies, plenty of new applications and tools are released for public usage. Social media applications such as Instagram, Facebook and Snapchat have been gaining popularity due to how easy an individual can share their daily lives in the form of photographs or videos. These media shared can be revisited any time to allow the individual to relieve their memories of the past. However, as popular as these platforms may be, they do not provide a unique sense of reminiscence when browsing through their past media. These platforms lack the function to retrieve relevant results when there is an imperfect retrieval cue. As such, they are not able to relate media together as episodes in a person’s life and generate a memory playback which provides a sense of nostalgia. This project develops a collective memory system called SCSE Memory which caters to the community of the School of Computer Science and Engineering (SCSE) in Nanyang Technological University (NTU). SCSE Memory allows the retrieval of memories to be based on a community instead of any individual. This is achieved by extending and enhancing the current MyLife and SnapMemory applications. SnapMemory is an Android application that works in a similar fashion to the popular social media applications where individuals can annotate and upload media captured or stored on their smartphones. This SnapMemory application is modified to allow community usage with existing functions such as effective image capturing, annotation and uploading of media in the form of photograph or video. Individuals can now share their media to help build and contribute to a community memory. MyLife is a web application that does memory retrieval and playback based on search queries provided by a user. MyLife utilizes the Autobiographical Memory Adaptive Resonance Theory (AM-ART) network model to emulate the autobiographical memory in human beings. The usage of this computational neural network model allows MyLife to handle imperfect queries which most existing applications are not able to handle. It also supports memory wandering where the memory retrieval results might seem to be a random set but in fact are contextually related. The memory playback will be in the form of a slideshow where multiple effects such as playback speed, image filter and music will be used. This MyLife application is modified to cater to a community environment where search queries and memory retrieval results are based on the community. The mood vectors utilized for the memory retrieval are also modified to generate a more accurate result and allow better flexibility for future extensions. With all the enhancements and new functions in place, a comprehensive system has been formed to provide individuals with a sense of nostalgia and reminiscent when viewing past memories.