A method performed by an electronic device is described. The method includes receiving a set of images. The method also includes determining a motion region and a static region based on the set of images. The method further includes extracting, at a first rate, first features from the motion region. The method additionally includes extracting, at a second rate that is different from the first rate, second features from the static region. The method also includes caching the second features. The method further includes detecting at least one object based on at least a portion of the first features.