Classification performance of pattern recognition systems based on Cross-plot and artificial neural network.

Pre-classified information is based on Data Set A (60 targets) for testing Data Set B (60 targets). Then, the combination of Data Set A and B (total 120 targets) is for testing Data Set C (60 positive targets) based on extraction of one target from every 5th frame of the video that consists of a total of 304 frames. Here, D1 and D2 are defined as the pre-classified and test data sets. The PTP PFP and PFN are defined as the True-Positive, False-Positive, and False-Negative result classification. The Precision and Recall are equal to PTP over (PTP + PFP) and PTP over (PTP + PFN) respectively. Since Data Set C comprises of the same air target class and established as all positive target outputs, the False-Positive result classification is not applicable. Both Precision and Recall results are higher for the Cross-plot technique in comparison to its artificial neural network counterpart.