Search

SIFT (Scale Invariant Feature Transform) generates image features widely used to match objects in different images. Previous work on hardware-based SIFT implementation requires excessive internal memory and hardware logic [1]. In this paper, a new hardware organization is proposed to implement the keypoint detection in SIFT with a less memory and hardware cost than the previous work. To this end, a parallel Gaussian filter bank is adopted to eliminate the buffers that store intermediate results because parallel operations allow all intermediate results available at the same time. The processing order of the vertical filtering ahead of the horizontal operation also reduces the storage space. The computational complexity is reduced by sharing the Gaussian filter bank for multiple octaves. As a result, the memory size is reduced by more than 80 percent without a complexity increase in hardware design.

Hand posture classification is popular in systems that require an effective human-machine interface. Previous classification algorithms suffer from inaccurate results when it is difficult to distinguish a hand from a wrist. To overcome this difficulty, this paper proposes a new algorithm for hand posture classification that uses a histogram of convex defects around the segment of a hand to be classified. As the characteristics of convex defects do not vary significantly depending on inclusion of a wrist, the proposed algorithm does not suffer substantially from a reduced classification accuracy. Furthermore, the proposed algorithm is also rotation-invariant. Experimental results show that the correct classification ratio is 97.06% on average.

The SoC presented in this paper integrates an H.264 encoder with an ISP (Image Signal Processor). It is currently implemented in an FPGA and processes an HD-size (1280 × 720) image at the speed of 15 fps with the operating clock frequency of 50 MHz. In the presented demo system, a Bayer input from a CMOS image is given to the FPGA and the output stream is transmitted through an USB transceiver to a PC that decodes and displays the H.264 stream.