Detecting pedestrian has been arguably addressed as a special topic beyond general object detection. Although recent deep learning object detectors such as Fast/Faster R-CNN [1,2] have shown excellent performance for general object detection, they have limited success for detecting pedestrian, and previous leading pedestrian detectors were in general hybrid methods combining hand-crafted and deep convolutional features. In this paper, we investigate issues involving Faster R-CNN [2] for pedestrian detection. We discover that the Region Proposal Network (RPN) in Faster R-CNN indeed performs well as a stand-alone pedestrian detector, but surprisingly, the downstream classifier degrades the results. We argue that two reasons account for the unsatisfactory accuracy: (i) insufficient resolution of feature maps for handling small instances, and (ii) lack of any bootstrapping strategy for mining hard negative examples. Driven by these observations, we propose a very simple but effective baseline for pedestrian detection, using an RPN followed by boosted forests on shared, high-resolution convolutional feature maps. We comprehensively evaluate this method on several benchmarks (Caltech, INRIA, ETH, and KITTI), presenting competitive accuracy and good speed. Code will be made publicly available.

Sketch-based face recognition is an interesting task in vision and multimedia research, yet it is quite challenging due to the great difference between face photos and sketches. In this paper, we propose a novel approach for photo-sketch generation, aiming to automatically transform face photos into detail-preserving personal sketches. Unlike the traditional models synthesizing sketches based on a dictionary of exemplars, we develop a fully convolutional network to learn the end-to-end photo-sketch mapping. Our approach takes whole face photos as inputs and directly generates the corresponding sketch images with efficient inference and learning, in which the architecture is stacked by only convolutional kernels of very small sizes. To well capture the person identity during the photo-sketch transformation, we define our optimization objective in the form of joint generativediscriminative minimization. In particular, a discriminative regularization term is incorporated into the photo-sketch generation, enhancing the discriminability of the generated person sketches against other individuals. Extensive experiments on several standard benchmarks suggest that our approach outperforms other state-of-the-arts in both photosketch generation and face sketch verification.

Oxidative damage is involved in many chronic diseases including those cited as the major causes of death in Western societies such as cardiovascular disorders and cancer. Antioxidants may prevent these degenerative processes by various mechanisms including the scavenging of free radicals. Intake of antioxidant supplements is associated with preventing oxidative damages. This study investigated the absorption and antioxidant effects of a xanthone-rich mangosteen liquid in healthy human volunteers after the acute consumption of 59 mL of the supplement. The liquid contained mangosteen, aloe vera, green tea, and multivitamins. Results indicated that R-mangostin and vitamins B 2 and B 5 were bioavailable, with observed C max at t max of around 1 h. The antioxidant capacity measured with the oxygen radical absorbance capacity (ORAC) assay was increased with a maximum effect of 18% after 2 h, and the increased antioxidant level lasted at least 4 h. Overall, this study demonstrated the bioavailability of antioxidants from a xanthone-rich mangosteen product and its in vivo antioxidant effects.

No abstract

A novel high-throughput assay for measuring antioxidant capacity against superoxide anion has been developed and validated. In this assay, hydroethidine (HE), a fluorescent probe, is oxidized by superoxide anion generated by xanthine and xanthine oxidase and increases its fluorescence intensity. Therefore, the inhibition of loss of HE's fluorescence intensity in the presence of antioxidant is an index of antioxidant capacity. The result is expressed as superoxide dismutase (SOD) equivalent. Unlike other probes, such as tetrazolium and lucigenin, one major advantage of this assay is that the use of HE is not prone to artifact. The method was rigorously validated through linearity, precision, accuracy, and ruggedness. The linear range, limit of quantitation (LOQ), and limit of detection (LOD) are 0.22-3.75 units/mL, 0.30 unit/mL, and 0.10 unit/mL, respectively. A wide variety of phenolic compounds and fruit extracts were analyzed.