A 217W120fps Ultralowpower Dualmode Cmos Image Sensor With Senputing Architecture

• doi: 10.1109/ASP-DAC52403.2022.9712591
• title: A 2.17μW@120fps Ultra-Low-Power Dual-Mode CMOS Image Sensor with Senputing Architecture
• publisher: IEEE
• isbn: 978-1-6654-2136-2
• issn: 2153-6961
• rank: 342
• access_type: LOCKED
• content_type: Conferences
• abstract: This paper proposes an ultra-low-power CMOS Image Sensor (CIS) chip based on sensing-with-computing (Senputing) architecture to reduce the power bottleneck of vision system. This Senputing chip achieves BNN 1st-layer convolution in analog domain with ultra-low power consumption. It has two working modes, Normal-Sensor (NS) mode and Direct- Photocurrent-Computation (DPC) mode. The prototype measurement results under 65nm CMOS process on MNIST classification task shows that the power of feature map computation is 2.17μW with 120fps frame rates and 98.1% accuracy. The computation efficiency reaches to 11.49TOPs/W, which is 14.8× higher than state-of-art works.
• article_number: 9712591
• pdf_url: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9712591
• html_url: https://ieeexplore.ieee.org/document/9712591/
• abstract_url: https://ieeexplore.ieee.org/document/9712591/
• publication_title: 2022 27th Asia and South Pacific Design Automation Conference (ASP-DAC)
• conference_location: Taipei, Taiwan
• conference_dates: 17-20 Jan. 2022
• publication_number: 9712466
• is_number: 9712479
• publication_year: 2022
• publication_date: 17-20 Jan. 2022
• start_page: 92
• end_page: 93
• citing_paper_count: 1
• citing_patent_count: 0
• download_count: 718
• insert_date: 20220221
• index_terms:
  • ieee_terms:
    • Semiconductor device measurement
    • Power measurement
    • Design automation
    • Machine vision
    • Prototypes
    • Computer architecture
    • CMOS image sensors
  • dynamic_index_terms:
    • Power Consumption
    • Feature Maps
    • Analog Domain
    • Convolutional Neural Network
    • Photodiode
    • Focal Plane
    • Back Focal Plane
    • Hierarchical System
    • Direct Computation
    • MNIST Dataset
    • MNIST Data Set
    • Shift Register
• isbn_formats:
  • format: Print on Demand(PoD) ISBN, value: 978-1-6654-2136-2, isbnType: New-2005
  • format: USB ISBN, value: 978-1-6654-2134-8, isbnType: New-2005
  • format: Electronic ISBN, value: 978-1-6654-2135-5, isbnType: New-2005
• authors:
  • Author Name: Ziwei Li
    Affiliation: Beijing Jiaotong University, China
    Author URL: https://ieeexplore.ieee.org/author/37089267970
    ID: 37089267970
    Order: 1
    Author Affiliations:
    • Beijing Jiaotong University, China
  • Author Name: Han Xu
    Affiliation: Tsinghua University, China
    Author URL: https://ieeexplore.ieee.org/author/37086593786
    ID: 37086593786
    Order: 2
    Author Affiliations:
    • Tsinghua University, China
  • Author Name: Zheyu Liu
    Affiliation: Tsinghua University, China
    Author URL: https://ieeexplore.ieee.org/author/37085778930
    ID: 37085778930
    Order: 3
    Author Affiliations:
    • Tsinghua University, China
  • Author Name: Li Luo
    Affiliation: Beijing Jiaotong University, China
    Author URL: https://ieeexplore.ieee.org/author/37085384804
    ID: 37085384804
    Order: 4
    Author Affiliations:
    • Beijing Jiaotong University, China
  • Author Name: Qi Wei
    Affiliation: Tsinghua University, China
    Author URL: https://ieeexplore.ieee.org/author/37582779500
    ID: 37582779500
    Order: 5
    Author Affiliations:
    • Tsinghua University, China
  • Author Name: Fei Qiao
    Affiliation: Tsinghua University, China
    Author URL: https://ieeexplore.ieee.org/author/37272195200
    ID: 37272195200
    Order: 6
    Author Affiliations:
    • Tsinghua University, China

Image Sensor

• sensor_type: CMOS
• resolution: 32×32 pixels
• dynamic_range: Not specified
• pixel_size: Not specified
• dark_current: Not specified

Optical Data

• focal_length: Not specified
• aperture: Not specified
• field_of_view: Not specified
• distortion: Not specified

Performance Metrics

• frame_rate: 120 fps
• power_consumption: 2.17μW

Applications & Benefits

• cell_imaging: The sensor works on MNIST classification tasks for pattern recognition in images.
• benefits: Ultra-low power consumption (2.17μW) with high computational efficiency (11.49TOPs/W) and an accuracy of 98.1% on the MNIST dataset.

Supporting Organizations

• supported_by: Beijing Jiaotong University, Tsinghua University

Manuscript Details

• publication_date: 17-20 Jan. 2022

Relevancy Score

• score: 9
• missing_fields:
  • dynamic_range
  • pixel_size
  • dark_current
  • focal_length
  • aperture
  • field_of_view
  • distortion
  • signal_to_noise_ratio
  • sensitivity
  • shutter_speed
  • noise

Processed JSON Filename

request_239d0bfe-c96b-4888-960c-efbd73119dd0-a_217w120fps_ultralowpower_dualmode_cmos_image_sensor_with_senputing_architecture.json