Systems and methods are disclosed for object detection by receiving an image; segmenting the image; extracting features from the image; and performing a dimension-wise spatial layout selection to pick up dimensions inside a discriminative spatial region for classification.
展开▼