An efficient face segmentation algorithm based on binary partition tree

This paper presents an efficient face segmentation algorithm based on binary partition tree. Skin-like regions are first obtained by integrating the results of pixel classification and watershed segmentation. Facial features are extracted by the techniques of valley detection and entropic thresholding, and are used to refine the skin-like regions. In order to segment the facial regions from the skin-like regions, a novel region merging algorithm is proposed by considering the impact of the common border ratio between adjacent regions, and the binary partition tree is used to represent the whole region merging process. Then the facial likeness of each node in the binary partition tree is evaluated using a set of fuzzy membership functions devised for a number of facial primitives of geometrical, elliptical and facial features. Finally, an efficient algorithm of node selecting in the binary partition tree is proposed for the final face segmentation, which can exactly segment the faces without any underlying assumption. The performance of the proposed face segmentation algorithm is demonstrated by experimental results carried out on a variety of images in different scenarios.     

An improved ultrasound‐assisted extraction process of gossypol acetic acid from cottonseed soapstock

To investigate the extracted process of gossypol acetic acid (G‐AA) from cottonseed soapstock and explore the improvement of its yield and purity, a novel ultrasound‐assisted extraction and crystallization method was introduced to this process. Under the optimized conditions, preliminary G‐AA with the yield of 1300 mg and the purity of 95.9% could be obtained from 100 g of fresh soapstock by ultrasound‐assisted extraction. In addition, UV, IR, and NMR spectrum further confirmed the detailed chemical structure of G‐AA. Assay of inhibiting human prostate tumor cell line PC‐3 and human breast cancer cell line MDA‐MB‐231 revealed its biological activity, the values of IC₅₀ are 9.096 μmol/L and 14.37 μmol/L respectively. In comparison with the conventional solvent extraction, this novel process increases the content of G‐AA over 90%, reduces the time of crystallization by 75%, and retains the anticancer activity of gossypol. © 2009 American Institute of Chemical Engineers AIChE J, 2009     

Strain-induced internal fibrillation in looped aramid filaments

By mapping the small-angle X-ray scattering (SAXS) from a looped poly-(para-phenylene terephtalamide) (aramid, PPTA) filament using a synchrotron X-ray microbeam, we investigate the effects of axially compressive and tensile strain on internal fibrillar structures. Unique observations of oscillations in the highly anisotropic SAXS patterns indicate a regular internal structure. Upon increase of the applied compressive strain, a significant decrease in oscillation frequency is observed in the scattering pattern. With an increase in imposed tensile strain an increase in oscillation frequency in the scattering pattern is observed. One model capable of describing the intensity is a model of stacked cylinders. These cylinders could be part of the fibrillar structure present in the PPTA fibres, which consist of cylinder- to tape-like objects, the presence of which is supported by SEM images. One hypothetical physical interpretation presented here for the appearance of a regular internal structure is the occurrence of fibrillar separation in regions undergoing axially compressive strain, and the appearance of strain relief/slip planes between packs of fibrils in regions undergoing tensile strain. Compaction of these packs upon increase of tensile strain could increase the repetition distance in the structure as multiple packs combine to form a single pack.     

Vessel segmentation and width estimation in retinal images using multiscale production of matched filter responses

Automated segmentation of blood vessels in retinal images can help ophthalmologists screen larger populations for vessel abnormalities. However, automated vessel extraction is difficult due to the fact that the width of retinal vessels can vary from very large to very small, and that the local contrast of vessels is unstable. Further, the small vessels are overwhelmed by Gaussian-like noises. Therefore the accurate segmentation and width estimation of small vessels are very challenging. In this paper, we propose a simple and efficient multiscale vessel extraction scheme by multiplying the responses of matched filters at three scales. Since the vessel structures will have relatively strong responses to the matched filters at different scales but the background noises will not, scale production could further enhance vessels while suppressing noise. After appropriate selection of scale parameters and appropriate normalization of filter responses, the filter responses are then extracted and fused in the scale production domain. The experimental results demonstrate that the proposed method works well for accurately segmenting vessels with good width estimation.