Neuro-Inspired Vision approach, based on models from biology, allows to reduce the computational complexity. One of these models - The Hmax model - shows that the recognition of an object in the visual cortex mobilizes V1, V2 and V4 areas. From the computational point of view, V1 corresponds to the area of the directional filters (for example Sobel filters, Gabor filters or wavelet filters). This information is then processed in the area V2 in order to obtain local maxima. This new information is then sent to an artificial neural network. This neural processing module corresponds to area V4 of the visual cortex and is intended to categorize objects present in the scene. In order to realize autonomous vision systems (consumption of a few milliwatts) with such treatments inside, we studied and realized in 0.35μm CMOS technology prototypes of two image sensors in order to achieve the V1 and V2 processing of Hmax model.
Today, solid state image sensors are used in many applications like in mobile phones, video surveillance
systems, embedded medical imaging and industrial vision systems. These image sensors require the integration
in the focal plane (or near the focal plane) of complex image processing algorithms. Such devices must meet the
constraints related to the quality of acquired images, speed and performance of embedded processing, as well
as low power consumption. To achieve these objectives, low-level analog processing allows extracting the useful
information in the scene directly. For example, edge detection step followed by a local maxima extraction will
facilitate the high-level processing like objects pattern recognition in a visual scene. Our goal was to design an
intelligent image sensor prototype achieving high-speed image acquisition and non-linear image processing (like
local minima and maxima calculations). For this purpose, we present in this article the design and test of a 64×64
pixels image sensor built in a standard CMOS Technology 0.35 μm including non-linear image processing. The
architecture of our sensor, named nLiRIC (non-Linear Rapid Image Capture), is based on the implementation of
an analog Minima/Maxima Unit. This MMU calculates the minimum and maximum values (non-linear functions),
in real time, in a 2×2 pixels neighbourhood. Each MMU needs 52 transistors and the pitch of one pixel is
40×40 mu m. The total area of the 64×64 pixels is 12.5mm2. Our tests have shown the validity of the main functions
of our new image sensor like fast image acquisition (10K frames per second), minima/maxima calculations in
less then one ms.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.