VCI2R at the NTCIR-13 Lifelog-2 Lifelog Semantic Access Task

VCI2R at the NTCIR-13 Lifelog-2 Lifelog Semantic Access Task
Title:
VCI2R at the NTCIR-13 Lifelog-2 Lifelog Semantic Access Task
Other Titles:
The 13th NTCIR Conference on Evaluation of Information Access Technologies
DOI:
N/A
Publication Date:
08 December 2017
Citation:
Abstract:
In this paper we describe our automatic approach for the NTCIR-13 Lifelog Semantic Access Task. The task is to query relevant lifelog images from user's daily life given an event topic. A major challenge is how to bridge the semantic gap between lifelog images and event-level topics. We propose a general framework to address this problem, with key components of various CNNs to translate lifelog images to object and scene features, relevant object/scene concepts searching for events, feature weighting adapted to events, and temporal smoothing to incorporate semantic coherence into the similarity between each image and query event. We achieved an o cial result 57.6% in terms of mean precision over 20 topics. We also analyze the effect of key components to the retrieval system.
License type:
PublisherCopyrights
Funding Info:
A*STAR JCO VIP REVIVE Project (1335h0009)
Description:
ISBN:

Files uploaded:

File Size Format Action
04-ntcir13-lifelog-linj.pdf 674.13 KB PDF Open