SIMILAR INSPECTION DATA RETRIEVAL FOR ROAD STRUCTURE INSPECTION BASED ON CANONICAL CORRELATION BETWEEN EYE TRACKING DATA AND INSPECTION RECORDS

Abstract

 This paper presents a retrieval method of similar inspection records using inspectors’ eye tracking data for supporting more efficient and advanced maintenance of road structures such as bridges and tunnels. In the proposed method, inspectors’ eye tracking data for distress images are obtained, and Canonical Correlation Analysis (CCA) between features extracted from eye tracking data and inspection records is performed. CCA gives projections which maximize correlation between those features and transforms them into new feature spaces. The similarities between inspection records can be calculated based on the transformed features. Furthermore, the obtained projections enable calculation of features based on inspectors’ eye tracking data even if new eye tracking data are not obtained as in future inspection. Then the proposed method provides inspection records similar to input records on the basis of inspectors’ eye tracking data. Therefore, inspectors can evaluate distresses referring to past similar records retrieved by the proposed method. As a result, the proposed method can expect more efficient and advanced maintenance inspection. In the last of this paper, the experiments with real data verify the effectiveness of the proposed method.