A knowledge engineering approach for image classification based on probabilistic reasoning systems
Seungyup Paek, Shih‐Fu Chang · 2002
We present an image classification system based on probabilistic reasoning that was developed to meet two research objectives. The first objective was to exploit interactions between multiple classifiers to improve classification performance. The second objective was to exploit existing descriptions of an image to improve classification performance. A case study of applying this approach for the consumer photograph domain is presented, along with performance evaluation results.