Fast Extraction of Objects of Interest from Images with Low Depth of Field

Abstract

In this paper, we propose a novel unsupervised video object extraction algorithm for individual images or image sequences with low depth of field (DOF). Low DOF is a popular photographic technique which enables the representation of the photographer's intention by giving a clear focus only on an object of interest (OOI). We first describe a fast and efficient scheme for extracting OOIs from individual low‐DOF images and then extend it to deal with image sequences with low DOF in the next part. The basic algorithm unfolds into three modules. In the first module, a higher‐order statistics map, which represents the spatial distribution of the high‐frequency components, is obtained from an input low‐DOF image. The second module locates the block‐based OOI for further processing. Using the block‐based OOI, the final OOI is obtained with pixel‐level accuracy. We also present an algorithm to extend the extraction scheme to image sequences with low DOF. The proposed system does not require any user assistance to determine the initial OOI. This is possible due to the use of low‐DOF images. The experimental results indicate that the proposed algorithm can serve as an effective tool for applications, such as 2D to 3D and photorealistic video scene generation.

Segmentation of Depth-of-Field Images Based on the Response of ICA Filters

“…4, it is easy to see that the figure on the left column has three main areas in different degrees of focus. The segmentation result of the HOS map from [1] and [2] is noisy providing only an almost binary characterization of the scene. The map P represents the energy at the frequency bands of the ICA filters and seems to only discriminate what is most in focus.…”

“…The first method is the higher-order-statistics map (HOS) used by the researches [1] and [2]. This map is calculated based on the fourth-order moment of the neighborhoods x i j , i.e., …”

“…Here, additional information such as color data [2] and further post processing techniques such as region merging [1] are not used. The results demonstrate, however, that the proposed method is extremely competitive generating smooth and noiseless segmented images.…”

“…Segmentation of focused objects is generally achieved by firstly computing a map which highlights structural complexity in the image [1], [2]. In referenced works, this map is then post-processed or integrated with other clues such as color information to improve results.…”

Automatic segmentation of focused objects from images with low depth of field

“…Different statistical features, such as local variance (Won et al, 2002) and forth-order moments (Kim, 2005) in the spatial domain, the variance (Wang et al, 2001) and the multi-scale description (Ye and Lu, 2002) of high-frequency coefficients in the wavelet domain, are exploited to measure the degree of focus for each pixel. The method in (Kim et al, 2007) is further extended to block based processing for speedup. However, the above region based methods usually require a sufficiently blurred background for a reliable segmentation of focused objects.…”