Previous studies (eg Kehrer, 1989 Spatial Vision4 45 – 62; Gurnsey et al, 1996 Journal of Experimental Psychology: Human Perception and Performance22 738 – 757) have shown that performance peaks several degrees from fixation in texture segmentation tasks, and performance falls as the target texture moves closer to the fovea or further into the periphery. There are two theories for this phenomenon: (1) neural processing speed in the fovea is slower than in the periphery (Kehrer 1989), and (2) the spatial frequency band of the texture is too low (ie too coarse) for the foveal receptive fields (Gurnsey et al 1996). However, the use of backward masking in previous studies made it impossible to decide between the two factors. The purpose of the present study was to isolate them. In experiment 1 a new stimulus configuration with backward masking was used, and previous reports were replicated. In experiment 2, the same texture was presented for 110 ms without a mask, but with added random-dot noise. Without limitations on processing time, the mid-peripheral advantage disappeared, which indicated that the previous findings were due to slower neural processing in the fovea. In experiment 3, a new type of texture was devised consisting of pairs of vertical lines with a horizontal offset. The offset was reversed for the target. Performance for unmasked 110 ms presentation was worst near the fovea and improved monotonically up to 12 deg. This peripheral advantage was spatial, not temporal, because it arose from larger receptive field sizes in periphery. When these results are taken together, the present study demonstrates that there are two independent causes for the mid-peripheral advantage in texture segregation.
We propose an algorithm to detect characters in images . The proposed algorithm consists ofthree phases ; training , candidate area detection , and character detection . In training Phase, a classifier to detect the ch 日racter is generated fセ om several haar − like feature − based weak classifiers by using Adaboost. In candidate area detection phase, one h馳ar − like feature is used to de丘ne the area to search charac 重 ers . And in character detection phase, the generated Adaboost − based classi 行er is used to detect the character in the defined area . We perfbrmed preliminary experiments to evaluate the proposed algorithm using 13 irnages . The experimental results show that some character en homogeneQus b日ckground can be
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.