Abstract: The “semantic gap” in image indexing and retrieval refers to the discontinuity between the low-level features that can be currently computed automatically (like color, shape and texture) and the richness of semantics in user queries posed for media search and retrieval. In this context, my talk will first present the different facets of the “semantic gap” problem and expose then actual research efforts and results towards the bridging of the gap, e.g., mid-level features or domain context-information. In this line, I will show how MPEG-7 based multimedia indexing and retrieval techniques enable the extraction of botanical characteristics for semi-automatic plant recognition. The identification of plants requires recognition and determination of plant species by using specific mid-level features such as bloom color, inflorescences, shape of blooms, number of petals and shape of leaves.
Abstract: The “semantic gap” in image indexing and retrieval refers to the discontinuity between the low-level features that can be currently computed automatically (like color, shape and texture) and the richness of semantics in user queries posed for media search and retrieval. In this context, my talk will first present the different facets of the “semantic gap” problem and expose then actual research efforts and results towards the bridging of the gap, e.g., mid-level features or domain context-information. In this line, I will show how MPEG-7 based multimedia indexing and retrieval techniques enable the extraction of botanical characteristics for semi-automatic plant recognition. The identification of plants requires recognition and determination of plant species by using specific mid-level features such as bloom color, inflorescences, shape of blooms, number of petals and shape of leaves.