Visual‐based and object‐conscious image retrieval by block reallocation into object region

Abstract

Visual-based' image retrieval based on the visual similarities over the entire image is one of the powerful and useful ways when targeting large volume content with inadequate annotation. Generally, conventional methods divide a query image and target images in the database into grid-shaped blocks and calculate the similarity based on image features by comparing each corresponding block straightforwardly. However, the method sometimes fails in terms of object-conscious retrieval when their backgrounds are almost the same but the only the object is different or the object's size and/or position is different. To solve the problem, we propose a new method featuring the reallocation of some blocks to the object region (OB-blocks) and a new similarity score by placing a weight on the OB-blocks, which are derived from the visual saliency map. Our proposed method could realize 'visual-based and object-conscious' image retrieval. We verify the effectiveness of this method through comparison experiments.

Non-rigid Interest Object Extraction Using Minimum Convex Coverage

“…In the content-base category, a lesion region in a radiograph, CT, or MRI gray image is commonly described by a rectangle ROI, which center and diameter (or diagonal) were estimated by unsupervised image processing methods, e.g., histogram thresholding, image transformation, pixel clustering or template-matching [15,23,19]. The regionbased category aimed to improve region-matching performance to the desired interest object, in which some region saliency strategies [36], e.g., attention window or supervision information [12,33,28], were incorporated into ROI selection by manual or semi-manual [21,2]. From then on, especially at the increasing requirements of interest object annotation, retrieval, detection and positioning, ROI selection and pixel extraction were extended to a wider scope of applications, e.g., tricolor natural images or multi-spectral RS images, even together with complex backgrounds.…”