A 0.8\mu\mathrm{m}$ 32Mpixel Always-On CMOS Image Sensor with Windmill-Pattern Edge Extraction and On-Chip DNN
This paper presents a CMOS image sensor (CIS) that integrates two operation modes: a high-resolution viewing mode with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0.8 \mu \mathrm{m}$</tex> 32Mpixels and a low-power always-on object recognition mode consuming 2.67 mW at 10 fps. The CIS features a unique windmill-pattern analog edge extraction circuit that is resilient to illumination variations. An on-chip DNN processor was implemented alongside a compact algorithm with only 12 KB for coefficients and 48 KB for working memory. The design incorporates separate circuit areas for high-speed viewing and low-power sensing modes, thereby ensuring optimal performance and energy efficiency.
Authors
- Kazutaka Yasuda
- Ota K (ORCID: https://orcid.org/0000-0002-5628-9662)
- C. Takano
- Kazutoshi Haraguchi (ORCID: https://orcid.org/0000-0003-0919-3024)
- Keiji Matsuda (ORCID: https://orcid.org/0000-0002-1507-1077)
- Setsuo Minami
- M. Tsuruta
- K. Yamanaka
- Motoyuki Sato
- Yosuke Ueno (ORCID: https://orcid.org/0000-0002-0402-9914)
- Etsushi Kato (ORCID: https://orcid.org/0000-0001-8814-804X)
- Garuda FUJII (ORCID: https://orcid.org/0000-0002-8867-6544)
- Makoto Watanabe (ORCID: https://orcid.org/0000-0002-6147-7713)
- Akihiko Kato (ORCID: https://orcid.org/0000-0001-5556-3845)
- K. Yasuoka
- Katsurou Hanzawa
- Sachio Akebono
Publication Details
- Published
- 2025-06-08
- DOI
- https://doi.org/10.23919/vlsitechnologyandcir65189.2025.11074825
- Primary Topic
- CCD and CMOS Imaging Sensors
- Type
- article
- Field-Weighted Citation Impact
- 0.00