SalientShape: group saliency in image collections

Abstract

Efficiently identifying salient objects in large image collections is essential for many applications including image retrieval, surveillance, image annotation, and object recognition. We propose a simple, fast, and effective algorithm for locating and segmenting salient objects by analysing image collections. As a key novelty, we introduce group saliency to achieve superior unsupervised salient object segmentation by extracting salient objects (in collections of pre-filtered images) that maximize between-image similarities and within-image distinctness. To evaluate our method, we construct a large benchmark dataset consisting of 15 K images across multiple categories with 6000+ pixel-accurate ground truth annotations for salient object regions where applicable. In all our tests, group saliency consistently outperforms state-of-the-art single-image saliency algorithms, resulting in both higher precision and better recall. Our algorithm successfully handles image collections, of an order larger than any existing benchmark datasets, consisting of diverse and heterogeneous images from various internet sources.

Content based image retrieval using salient orientation histograms

“…4. F β evaluation results for THUR15K database for different segmentation methods: FT [7], SEG [6], RC [4], GS [13] and the proposed method.…”

“…The THUR15000 database was introduced in [13] as the largest dataset for content based image retrieval using randomly selected internet images, downloaded from Flickr for 5 keywords: "butterfly", "coffee mug", "dog jump", "giraffe" and "plane". The database also contains salient regions marked at pixel accuracy, for images where such region exists.…”

“…In the first part, the proposed SOH based SBIR system is compared to Flickr and the SHOG based SBIR [3] (retrieval results are given in [13] for the best 50 and the best 100 retrieved images for each category. An image is considered as true positive if it contains a target object specified by the keywords.…”

“…In the second part, the images with pixel accurate ground truth are segmented, by applying the proposed automatic saliency-based initialization. The segmentation method is compared to other state-of-the-art techniques: FT [7], SEG [6], RC [4] and GS [13]. To weigh precision more than recall (advised in [7]), the F β is calculated with β 2 = 0.3: Figure 4 shows that the proposed automatic detection technique is able to perform at high accuracy, achieving the highest performance in the majority of the keywords and for the average F β (Table 2) on the whole database.…”

“…A salient orientation histogram (SOH) is built using the salient keypoints and the improved edge map. The proposed method is tested on the THUR15000 database [13] for image retrieval and detection purposes.…”