Replicate Statistics for Efficient Vision System Evaluation
نویسندگان
چکیده
Computer vision system evaluation faces problems stemming from high-dimensionality, difficulty in achieving statistical soundness, and genericity. In this paper, a new framework for computer vision system evaluation is introduced. This framework, when applied properly, can be used to evaluate, with statistical soundness, many different vision system components including scenes, sensors, and algorithms. In addition, it is shown how replicate statistics can be applied to obtain confidence intervals of the evaluation estimates and significantly reduce the amount of data required to reach such sound evaluatory conclusions. Included is a sample case study showing how the framework may be applied to evaluate sensor effects on a face recognition system.
منابع مشابه
Efficient Evaluation of Classification and Recognition Systems
In this paper, a new framework for evaluating a variety of computer vision systems and components is introduced. This framework is particularly well suited for domains such as classification or recognition systems, where blind application of the i.i.d. assumption would reduce an evaulation’s accuracy, such as with classification or recognition systems. With a few exceptions, much of the previou...
متن کاملModeling the efficiency of banks by cover data method and Genetic programming
Performance assessment of organizations in making strategic decisions for the future is essential. Thus, the efficiency determinations of organizations through planned economic growth in the future system loads are necessary. One of the most important indicators of performance evaluation is measuring efficiency. Banks are one of the most important economic units that provide proper conditions f...
متن کاملCost-Effective Video Filtering Solution for Real-Time Vision Systems
This paper presents an efficient video filtering scheme and its implementation in a field-programmable logic device (FPLD). Since the proposed nonlinear, spatiotemporal filtering scheme is based on order statistics, its efficient implementation benefits from a bit-serial realization. The utilization of both the spatial and temporal correlation characteristics of the processed video significantl...
متن کاملAn evaluation of the replicate pool method: quick estimation of genome-wide linkage peak p-values.
The calculation of empirical p-values for genome-wide non-parametric linkage tests continues to present significant computational challenges for many complex disease mapping studies. The gold standard approach is to use gene dropping to simulate null genome scans. Unfortunately, this approach is too computationally expensive for many data sets of interest. An alternative, more efficient method ...
متن کاملPose Evaluation Based on Bayesian classification Error
Pose evaluation is a fundamental issue in image processing and computer vision. In this paper, we propose a new method called BCE for pose evaluation based on Bayesian classification error. Various image cues are incorporated to depict an object including object shape, side region statistics and temporal information. Then a PEF (Pose Evaluation Function) is constructed based on Bayesian classif...
متن کامل