We present the MemoriEase retrieval system used for our participation in the NTCIR Lifelog-5 Task. We report our method to address the lifelog retrieval problem and discuss our official results of the MemoriEase at Lifelog-5 task. We originally introduced the MemoriEase system for the Lifelog Search Challenge (LSC) as an
interactive lifelog retrieval system. We have modified it to an automatic retrieval system to address the NTCIR Lifelog-5 Task. We
propose the BLIP-2 model as the core embedding model to retrieve
lifelog images from textual queries. The open-sourced Elasticsearch
search engine serves as the main engine in the MemoriEase system.
Some pre-processing and post-processing techniques are applied to
adapt this system to an automatic version and improve the accuracy
of retrieval results. Finally, we discuss the results of the system on
the task, some limitations of the system, and lessons learned from
participating in the Lifelog-5 task for further improvements for the
system in the future.