Consumer Document Analytical Accelerator Hardware

Document scanning devices are used for visual character recognition, followed by text analytics in the software. Often such extraction is insecure, and any third party can manipulate information. On other hand, near-edge processing restrained limited resources connectivity issues. The primary factors that lead to exploring independent hardware with natural language (NLP) capabilities latency during cloud computing costs. This paper introduces a accelerator information retrieval using memristive TF-IDF implementation. In this system, each sentence represented crossbar layer, column containing single word. number of matching scores TF IDF values was implemented operational amplifier-based comparator accumulator circuits. circuit designed 180nm CMOS process, Knowm Multi-Stable Switch memristor model, WOx device parameters. We compared its performance standard benchmark dataset. Variability device-to-device related issues were also taken into consideration analysis. concludes implementing score calculation applications as summarization.