Portal de Periódicos da CAPES

Bit-Plane Extracted Moving-Object Detection Using Memristive Crossbar-CAM Arrays for Edge Computing Image Devices

2018; Institute of Electrical and Electronics Engineers; Volume: 6; Linguagem: Inglês

10.1109/access.2018.2819986

2169-3536

Nazgul Dastanova, Sultan Duisenbay, Olga Krestinskaya, Alex Pappachen James,

Visual Attention and Saliency Detection

In this paper, we present the hardware implementation of a novel algorithm for moving-object detection, which can be integrated with CMOS image sensors. Bit planes of consecutive frames are stored in memristive crossbar arrays and compared using threshold-logic XOR gates. The resulting outputs are combined using weighted summation circuits and thresholded using comparators, to obtain binary images. A resistive content-addressable memory (CAM) array is used in the output stage to observe the numbers of different object pixels in the first and second pairs of the processed frames, in a row-by-row manner. The CAM array output conveys information on the motion direction and allows for optimal memory utilization through the selective row-wise storage of different bits. The proposed method outperforms the conventional moving-object detection algorithms, in terms of accuracy, specificity, and positive prediction metrics, and performs comparably in terms of other metrics.