Research of Face Location System Based on Human Vision Simulations

In this paper we present a robust face location system based on human vision simulations to automatically locate faces in color static images. Our method is divided into four stages. In the first stage we use a gauss low-pass filter to remove the fine information of images, which is useless in the initial stage of human vision. During the second and the third stages, our technique approximately detects the image regions, which may contain faces. During the fourth stage, the existence of faces in the selected regions is verified. Having combined the advantages of Bottom-Up Feature Based Methods and Appearance-Based Methods, our algorithm performs well in various images, including those with highly complex backgrounds.

A novel image restoration approach based on point location in high-dimensional space geometry

The goal of image restoration is to restore the original clear image from the existing blurred image without distortion as possible. A novel approach based on point location in high-dimensional space geometry method is proposed, which is quite different from the thought ways of existing traditional image restoration approaches. It is based on the high-dimensional space geometry method, which derives from the fact of the Principle of Homology-Continuity (PHC). Begin with the original blurred image, we get two further blurred images. Through the regressive deducing curve fitted by these three images, the first iterative deblured image could be obtained. This iterative "blurring-debluring-blurring" process is performed till reach the deblured image. Experiments have proved the availability of the proposed approach and achieved not only common image restoration but also blind image restoration which represents the majority of real problems.

adaptive location update mechanism for network-constrained moving objects in changeful traffic conditions

Natl Chiao Tung Univ, Dept Comp Sci

CHEMICAL-IONIZATION MASS-SPECTROMETRY OF SOME NITRO-CONTAINING BIFUNCTIONAL AROMATIC-COMPOUNDS - LOCATION OF PROTONATION SITE

It is found that the nitro substituent of some aromatic bifunctional compounds shows unusual reactivity towards protonation. In the chemical ionization mass spectra of nitrobenzoic acids and their esters and amides, and of nitrophenols and their ethers, protonations on the carboxyl, ester, amide, hydroxyl or alkoxyl groups are highly suppressed by that on the nitro group. As a result, fragmentations based on protonation on these groups unexpectedly become negligible. Ortho effects were observed for all the ortho isomers where the initial protonation on the nitro group is followed by an intramolecular proton transfer reaction, which leads to the expected 'normal' fragmentations. Protonation on the nitro substituent is much more favourable in energy than on any of the other substituents. The interaction of the two substituents through the conjugating benzene ring is found to be responsible for this 'unfair' competitive protonation. The electron-attracting nitro group strongly destabilizes the MH+ ions formed through protonation on the other substituent; although the COR (R = OH, OMe, OEt, NH2) groups are also electron-withdrawing, their effects are weaker than that of NO2; thus protonation on the latter group produces more-stable MH+ ions. On the other hand, an electron-releasing group OR (R = H, Me, Et) stabilizes the nitro-protonated species; the stronger the electron-donating effect of this group the more stable the nitro-protonated ions.

序列视觉信息整合的特性与机制

One of the great puzzles in the psychology of visual perception is that the visual world appears to be a coherent whole despite our viewing it through temporally discontinuous series of eye fixations. The investigators attempted to explain this puzzle from the perspective of sequential visual information integration. In recent years, investigators hypothesized that information maintained in the visual short-term memory (VSTM) could become visual mental images gradually during time delay in visual buffer and integrated with information perceived currently. Some elementary studies had been carried out to investigate the integration between VSTM and visual percepts, but further research is required to account for several questions on the spatial-temporal characteristics, information representation and mechanism of integrating sequential visual information. Based on the theory of similarity between visual mental image and visual perception, this research (including three studies) employed the temporal integration paradigm and empty cell localization task to further explore the spatial-temporal characteristics, information representation and mechanism of integrating sequential visual information (sequential arrays). The purpose of study 1 was to further explore the temporal characteristics of sequential visual information integration by examining the effects of encoding time of sequential stimuli on the integration of sequential visual information. The purpose of study 2 was to further explore the spatial characteristics of sequential visual information integration by investigating the effects of spatial characteristics change on the integration of sequential visual information. The purpose of study 3 was to explore the information representation of information maintained in the VSTM and integration mechanism in the process of integrating sequential visual information by employing the behavioral experiments and eye tracking technology. The results indicated that: (1) Sequential arrays could be integrated without strategic instruction. Increasing the duration of the first array could cause improvement in performance and increasing the duration of the second array could not improve the performance. Temporal correlation model was not fit to explain the sequential array integration under long-ISI conditions. (2) Stimuli complexity influenced not only the overall performance of sequential arrays but also the values of ISI at asymptotic level of performance. Sequential arrays still could be integrated when the spatial characteristics of sequential arrays changed. During ISI, constructing and manipulating of visual mental image of array 1 were two separate processing phases. (3) During integrating sequential arrays, people represented the pattern constituted by the objects' image maintained in the VSTM and the topological characteristics of the objects' image had some impact on fixation location. The image-perception integration hypothesis was supported when the number of dots in array 1 was less than empty cells, and the convert-and-compare hypothesis was supported when the number of the dot in array 1 was equal to or more than empty cells. These findings not only contribute to make people understand the process of sequential visual information integration better, but also have significant practical application in the design of visual interface.