Fig 1.
Presentation of image results over web search engines in semantically relevant titles/clusters.
Fig 2.
An overview of the proposed approach.
Fig 3.
Schematic representation of approach architecture depicting; Image results retrieval (a), Dense and complex graphs (b), Simplified image graphs (c), Converted image similarity graphs (d), Multimodal image tree (e), Image forest (f), and Exploration components (g).
Fig 4.
Query Formulation Panel (QFP) (a), Result Grid Panel (RGP) (b), Selection Panel (SP) (c), access of image result from the actual web source (d), Cluster List Panel (CLP) (e), Cluster-Graph Panel (CGP) (f) Graph Visualization Panel (GVP) (g), Information Window (h), and Related Results Panel (RRP) (i) of SUI.
Fig 5.
Cluster-graph data model construction over real image results.
Fig 6.
Schematic representation of user interaction paradigm; information need expression (a), the exploration initiation (b), linear navigation (c), and non-linear navigation (d).
Fig 7.
The grid exploration scenario (I) as a query is expressed in QFP (a), images results are presented in RGP (b), lookup of a selected image in SP (c), and access of an image from web source (d); The cluster-graph exploration scenario (II) as clusters hierarchy is presented in CLP (a), Cluster images presentation in RGP (b), lookup of a selected image in SP, exploration of an image in RRP (c) and lookup of a selected image in Information Window (d); The graph visualization exploration scenario (III) as images visualization in GVP (a), lookup of a selected image in Information Window (b), and access of an image from web source (c), against query ‘Bird’ AND ‘Red Bird’.
Fig 8.
Respondents task completion time of exploration tasks ET1 (a) and ET1 (b).
Fig 9.
Respondents time-on-panel in the performance of exploration tasks ET1 and ET2, (Where AP is the average precision of panels and Avg is the overall average).
Fig 10.
Respondents exploration activity in unstructured task ET2.
Table 1.
Respondents task-difficulty-ratings & user-experience evaluation.
Table 2.
Respondents SUS evaluation scores.
Table 3.
Respondents CUSQ evaluation scores.
Table 4.
Theoretical comparison of online image search tools; state-of-the-art search and exploration tools discussed in the literature and our proposed SUI approach.