high performance computing on graphics processing units: hgpu.org

We develop intermediate representations for deformable part models, and show that such representations have favorable performance characteristics for multi-class problems where the number of classes is large. Our model uses sparse coding of part filters to represent each filter as a sparse linear combination of shared dictionary elements. This leads to an universal set of parts that are shared among all object classes. Reconstruction of part responses via sparse-matrix-multiply reduces computation relative to conventional filter convolutions. Our model is well suited to a parallel implementation, and we report a new GPU DPM implementation which takes advantage of sparse coding of part filters. The speed-up offered by our intermediate representation and parallel computation enable real-time DPM detection of 20 different object classes on a laptop computer.