01-06-2013, 04:15 PM
Image Compression and Resizing for Retinal Implant in Bionic Eye
Bionic Eye.pdf (Size: 535.83 KB / Downloads: 26)
Abstract
One field where computer-related Image processing technology shows great promise for the future is
bionic implants such as Cochlear implants, Retinal implants etc.. Retinal implants are being developed
around the world in hopes of restoring useful vision for patients suffering from certain types of diseases
like Age-related Macular Degeneration (AMD) and Retinitis Pigmentosa (RP). In these diseases the
photoreceptor cells slowly degenerated, leading to blindness. However, many of the inner retinal
neurons that transmit signals from the photoreceptors to the brain are preserved to a large extent for a
prolonged period of time. The Retinal Prosthesis aims to provide partial vision by electrically
activating the remaining cells of the retina. The Epi retinal prosthesis system is composed of two
units, extra ocular unit and intraocular implant. The two units are connected by a telemetric inductive
link. The Extraocular unit consists of a CCD camera, an image processor, an encoder, and a transmitter
built on the eyeglass. High-resolution image from a CCD camera is reduced to lower resolution
matching the array of electrodes by image processor, which is then encoded into bit stream. Each
electrode in an implant corresponds to one pixel in an image. The bit stream is modulated on a 22 MHz
carrier and then transmitted wirelessly to the inside implant. This paper mainly discusses two
approaches in image processing which reduces the size of the image without loss of the object detection
rate to that of the original image. One is about the related image processing algorithms include image
resizing, color erasing, edge enhancement and edge detection. Second one is to generate the saliency
map for an image.
INTRODUCTION
Retinal Implant is a prosthetic device that maps visual images to control signals, based on
which it stimulates the surviving retinal circuitry. Image compression for bionic eye
compresses and resizes the images preserving the object detection rate of the image. The
resized image obtained has a comparable object detection rate to that of original image. This
indeed reduces the processing over head on implant inside the body. In eye the visual
information from the retina’s 130 million photoreceptors is compressed into electrical signals
carried by 1.2 million highly specialized ganglion neurons, whose axons form the optic nerve.
The optic nerve transmits visual information via the lateral geniculate nucleus to the primary
visual cortex of the brain. Blindness can result when any step of the optical pathway sustains
damage.
EPIRETINAL APPROACH
An Epiretinal prosthesis system usually employs a multi electrode array implanted on the
surface of the inner retina between the vitreous and internal limiting membrane. Epiretinal
prosthesis pass signal to the ganglion cells, while Subretinal prosthesis relay signals to the
bipolar cells. A data acquisition system located outside of the body captures images from the
surroundings, and converts the information into patterns of electrical signals. Upon the
reception of signals through data transmission and processing systems, the electrodes stimulate
the remaining retinal ganglion cells and restore vision[2]. Epiretinal approach is easier from
surgical point of view but mechanical anchoring of the implant to the epiretinal surface is
difficult.
METHODOLOGY
Retinal implant is a prosthetic device that maps visual images to control signals, based on
which it stimulates the surviving retinal circuitry. The epiretinal prosthesis system is composed
of two units, one extraocular and one intraocular [3]. The two units are connected by a
telemetric inductive link, allowing the intraocular unit to derive both power and a data signal
from the extraocular unit.
The extraocular unit includes a video camera and video processing board, a telemetry protocol
encoder chip, and an RF amplifier and primary coil. The intraocular unit consists of a
secondary coil, a rectifier and regulator, a retinal stimulator with a telemetry protocol decoder
and stimulus signal generator, and an electrode array.
At the extraocular unit side, a digital camera captures the image which is then preprocessed as
defined in sec(IV). The data is further processed by a pulse width modulation circuit (PWM)
and subsequently modulated onto an RF carrier using amplitude shift keying (ASK).
IInd APPROACH
Saliency maps are mainly used to get the salient parts of the image which will helpful for object
recognition in bionic eye[9]. So by using this method blind patients can have partial visibility of
the image. For generation of saliency map, first we have to generate image pyramids. The
image pyramid is a data structure designed to support efficient scaled convolution through
reduced image representation. It consists of a sequence of copies of an original image in which
both sample density and resolution are decreased in regular steps.
CONCLUSION
This paper discusses about the retinal implant architecture and extraocular image processing of
epiretinal prosthesis . To be able to meet the real-time requirements of a retinal prosthesis
system while running high-sophisticated image processing algorithms, some strategies for
transforming high resolution image to low resolution image, image edge enhancement and
detection algorithms have been presented. The image processing part in this paper is mainly
simulated in matlab and java. In future this work has to implement on suitable DSP processor.
The implant requirements have been outlined and the circuit design is explained in detail. A
major emphasis has been laid on the design for intraocular part —this makes the stimulator
ready for production and implantation.