Image Classification with SVM for CMOS Sensor Generating a Vector Image

In the coming Internet of Things (IoT) era, it is important to reduce volume data being output by sensors as well their power consumption. Since conventional image sensor for photography are often redundant in AI applications, that can lightweight use needed. this work, we propose a complementary metal oxide semiconductor (CMOS) (CIS) pixel circuit extract intensity gradients without analog memories. The available histogram oriented (HoG) features amount used classification. We performed experiments evaluate whether HoG calculated outputs from our pixels were suitable classification tasks. A support vector machine (SVM) classifier was trained with simulated human detection accuracy. also evaluated accuracy when quantized using "low bit decimation" and "value clipping" data. Our experimental results indicated highest 99.55% achieved 2-bit-width value clipping 4 × cells.