Toposequential effects on water balance and productivity in rainfed lowland rice ecosystem in Southern Laos

In the Mekong region, most paddies in rainfed lowland rice (Oryza sativa L.) lie in a sequence on gentle sloping land, and grain yield (GY) often depends on the toposequence position. There is, however, lack of information on toposequential effects on field water supply in rainfed lowland rice and how that influences GY. A total of eight field experiments were carried out on sandy, coarse-textured soils in Southern Laos (Champassak Province and Savannakhet Province) over three wet seasons (2000-2002). Components of the water balance, including downward and lateral water movement (D and L, respectively), were quantified at three different positions along toposequences (top, middle and bottom). GY, days-to-flower (DTF) and rainfall were measured, and the water productivity (WP) was determined. In most experiments, standing water disappeared first in the top position and gradually in lower positions. This was associated with the observation that when there was standing water in the field, the higher position had larger D in both the provinces and also larger L in Champassak Province. However, in one experiment, water loss appeared later in the higher position, as the result of lower L, apparently due to some water inputs other than rainfall occurring at this position. Despite larger D plus L at the top position, seasonal sum of D and L were not much affected by the toposequence position, as the daily rate of D plus L became minimal when the standing water was lost earlier in the top position. Lower GY was associated with earlier disappearance of standing water from the field. Relatively low GY was expected in the top toposequence position. This was clearly shown in the toposequence of Phonthong, Champassak Province, as the timing of standing water disappearance relative to flowering was earlier in the top position. Variation in GY across the toposequence positions was coupled with the WP variation, and both GY and WP tended to decline with increased DTF. Therefore, variation in productivity of rainfed lowland rice across toposequence positions depends mainly on the field water status around flowering time. (c) 2005 Elsevier B.V. All rights reserved.

Face recognition robust to head pose from one sample image

Most face recognition systems only work well under quite constrained environments. In particular, the illumination conditions, facial expressions and head pose must be tightly controlled for good recognition performance. In 2004, we proposed a new face recognition algorithm, Adaptive Principal Component Analysis (APCA) [4], which performs well against both lighting variation and expression change. But like other eigenface-derived face recognition algorithms, APCA only performs well with frontal face images. The work presented in this paper is an extension of our previous work to also accommodate variations in head pose. Following the approach of Cootes et al, we develop a face model and a rotation model which can be used to interpret facial features and synthesize realistic frontal face images when given a single novel face image. We use a Viola-Jones based face detector to detect the face in real-time and thus solve the initialization problem for our Active Appearance Model search. Experiments show that our approach can achieve good recognition rates on face images across a wide range of head poses. Indeed recognition rates are improved by up to a factor of 5 compared to standard PCA.