Modified model in content-based flower image retrieval

Flower image retrieval is a significant and challenging problem in content-based image retrieval. We had systematic and overall researches on flower images, including repetitive images filtering, regional segmentation, feature extraction and image retrieval based on SVM, etc. Firstly, in order to ensure retrieval results, we propose a repetitive images detection algorithm based on Canny edge to filter repetitive flower images. Aiming at image segmentation, we proposed an adaptive segmentation algorithm based on 2RGB mixed color model to segment flower images. On the basis of multi-feature fusion strategy, we propose a weighted invariant moment feature based on HSV color model to extract shape feature from flower images, and then we also propose an edge LBP operator which combine texture and shape information. Final experimental results on flower dataset reveal that our algorithms are effective.