Design and analysis of the human-computer interface for an image database.

摘要

Supporting effective retrieval of images from databases is a difficult problem as sound design techniques for facilitating retrieval are rare. The human-computer interface is the principal component responsible for facilitating retrieval in databases. Hence, image retrieval can be better facilitated by improving the HCI design. An HCI was created that included both visual and verbal means of image retrieval (Visual-Verbal). Another HCI was created, for comparison purposes, that only included verbal means of retrieval and displayed only verbal information (Verbal). An object-oriented environment and client/server design were used to develop the HCIs. A set of digitized moving images from popular films was used as the test data set. Applying both manual and automated techniques a practical yet powerful image indexing procedure was developed. The index contained both verbal and visual information and was fully integrated into the Visual-Verbal HCI as an online aid. Effectiveness of the two HCIs was compared using quantitative and qualitative data collected from search sessions involving 18 users. No significant difference was found, between the two HCIs, on the variables search completion, error, help request and optimum searching. However, in both HCIs search completion rate was high (more than 80%).