Vision-based Vehicle Type Classification Using Partial Gabor Filter Bank

A vision-based vehicle type classification method using partial Gabor filter bank is present in this paper for five vehicles categorization: sedan, van, hatchback sedan, bus and van truck. To reduce the influence caused by the hues of vehicles, we extract the Gabor features from the edge image of vehicle, instead of from the grey image. Partial Gabor filter bank approach, which can save memory and computation cost significantly, is introduced and a new partial sampling method is proposed. The experimental results show that the recognition rate reaches 95.17% using partial Gabor features, illustrating the effectiveness of the proposed approach.