科研成果 by Year: 2011

Water quality management is subject to large uncertainties due to inherent randomness in the natural system and vagueness in the decision-making process. For water quality management optimization models, this means that some model coefficients can be represented by probability distributions, while others can be expressed only by ranges. Interval linear programming (ILP) and risk explicit interval linear programming (REILP) models for optimal load reduction at the watershed scale are developed for the management of Lake Qionghai Watershed, China. The optimal solution space of an ILP model is represented using intervals corresponding to the lower and upper bounds of each decision variable. The REILP model extends the ILP model through introducing a risk function and aspiration levels (lambda(pre)) into the model formulation. The REILP model is able to generate practical solutions and trade-offs through solving a series of submodels, minimizing the risk function under different aspiration levels. This is illustrated in the present study by solving 11 submodels corresponding to different aspiration levels. The results show that the ILP model suffers severe limitations in practical decision support, while the REILP model can generate solutions explicitly relating system performance to risk level. Weighing the optimal solutions and corresponding risk factors, decision makers can develop an efficient and practical implementation plan based directly on the REILP solution.

General object recognition and image understanding is recognized as a dramatic goal for computer vision and multimedia retrieval. In spite of the great efforts devoted in the last two decades, it still remains an open problem. In this paper, we propose a selective attention-driven model for general image understanding, named GORIUM (general object recognition and image understanding model). The key idea of our model is to discover recurring visual objects by selective attention modeling and pairwise local invariant features matching on a large image set in an unsupervised manner. Towards this end, it can be formulated as a four-layer bottomup model, i.e., salient region detection, object segmentation, automatic object discovering and visual dictionary construction. By exploiting multi-task learning methods to model visual saliency simultaneously with the bottom-up and top-down factors, the lowest layer can effectively detect salient objects in an image. The second layer exploits a simple yet effective learning approach to generate two complementary maps from several raw saliency maps, which then can be utilized to segment the salient objects precisely from a complex scene. For the third layer, we have also implemented an unsupervised approach to automatically discover general objects from large image set by pairwise matching with local invariant features. Afterwards, visual dictionary construction can be implemented by using many state-of-the-art algorithms and tools available nowadays.