No abstract

Knowledge regarding the amount of blur perceived to be "bothersome" to an individual, namely that which is assumed to be annoying and to adversely affect task performance, remains limited. A Badal optical system was used to measure the blur detection, bothersome blur, and non-resolvable blur dioptric thresholds monocularly either to an isolated 20/50 or 20/200 Snellen E, or to three 20/50 lines of text. Subjects were comprised of 13 visually normal young adults and 3 absolute presbyopes. Cycloplegia was used to paralyze accommodation in the young adults. Within each target type for the young adults, the mean bothersome blur threshold was always significantly larger than that found for blur detection and significantly smaller than that found for non-resolvable blur. Across target types and blur criteria, the bothersome blur thresholds for the isolated 20/50 E (1.02 D) and the 20/50 text (1.34 D) were not significantly different, although in 12 of the 13 subjects the latter were larger (p<0.002, sign test). However, both were significantly smaller than for the isolated 20/200 E (1.80 D). In a subset of young adult subjects, bothersome blur was found to be repeatable over time. The results were similar in the absolute presbyopes. The bothersome blur threshold was primarily influenced by target detail and secondarily by target extent. These findings have important implications with respect to tolerances for optical lens design and refractive surgery outcomes, as well as provide insight into basic aspects of human blur perception.

The ACUVUE design was superior in stability for two of the four conditions tested. This resulted in a more stable lens immediately after insertion as well as during some visual tasks involving either naturally occurring or programmed large versional eye movements. Both lens designs provided acceptable performance in terms of induced astigmatism produced by off-axis rotation.

In this paper; we propose a novel scheme for automatic and fast detection of human faces in color images where the number; the location, the orientation and the size of the faces are unknown, under non-constrainedscene conditions such as complex background and uncontrolled illumination. First, each frame is segmented using skin chrominance values, providing face area candidates. Then, shape analysis and wavelet packet decomposition are performed on the face area candidates in order to detect human faces. Each face area candidate is described by a subset of bandjiltered images containing wavelet coeficients. These coeficients characterize the face texture and a set of simple statistical data is extracted in order to form compact and meaningful feature vectors. Then, an efJicient and reliable probabilistic metric derived from the Bhattacharrya distance is used in order to classify the face area candidate feature vectors into face or non-face areas.0-7695-0253-9/99 $10.00 0 1999 IEEE