Car image recognition with lacked input data

A car type recognition system, using methods of the R&FHPR/FF (Rough & Fuzzy Hierarchical Pattern Recognition using Fixation Feedback), infer just about correctly based on precise input data. Although, it is difficult that all features are extracted completely in general using image processing methods. And for the system used the fuzzy inference, if input datum is lacked, the system does not work well. In this paper, a frame work of a modified fuzzy inference method with lacked input data is introduced. And experimental results are provided. In the proposed method that is modified from the Mamdani's fuzzy inference, a result of each rule is the adjustment for the purpose of protection from influence of lacked input data. The adjustment of result fuzzy labels at each rule is used the degree of importance which is set up in the rules by human. In the results of experiment, the system can infer well with lacked input data used this method. In results of experiments used a set of eight fuzzy rules (five input and eight output), when all combination of all input data are lacked, the system infers so correctly.