Fast SIFT Design for Real-Time Visual Feature Extraction
Fast SIFT Design for Real-Time Visual Feature Extraction
复制标题
DOI：
10.1109/tip.2013.2259841
复制
发表时间：
2013-08-01
影响因子：
10.6
通讯作者：
Chang, Nelson Yen-Chung
中科院分区：
文献类型：
--
作者：
Chiu, Liang-Chi;Chang, Tian-Sheuan;Chang, Nelson Yen-Chung
使用尺度不变特征变换（SIFT）进行视觉特征提取广泛用于对象识别。然而，由于其帧级计算采用迭代高斯模糊操作，其实时实现存在延迟长、计算量大和内存存储量大的问题。因此，本文提出了一种具有积分图像的层并行 SIFT (LPSIFT)，及其并行硬件设计，具有满足实时应用需求的动态特征提取流程。与原始SIFT算法相比，该方法减少了90%的计算量和95%的内存占用。最终实现采用 580-K 门数和 90 nm CMOS 技术，并为 30 帧/秒的 VGA 图像提供 6000 个特征点/帧，类似于在 100 MHz 时钟速率下为 30 帧/秒的 1920x1080 图像提供 2000 个特征点/帧。
Visual feature extraction with scale invariant feature transform (SIFT) is widely used for object recognition. However, its real-time implementation suffers from long latency, heavy computation, and high memory storage because of its frame level computation with iterated Gaussian blur operations. Thus, this paper proposes a layer parallel SIFT (LPSIFT) with integral image, and its parallel hardware design with an on-the-fly feature extraction flow for real-time application needs. Compared with the original SIFT algorithm, the proposed approach reduces the computational amount by 90% and memory usage by 95%. The final implementation uses 580-K gate count with 90-nm CMOS technology, and offers 6000 feature points/frame for VGA images at 30 frames/s and similar to 2000 feature points/frame for 1920x1080 images at 30 frames/s at the clock rate of 100 MHz.