An image-sensor-based photo receiver for free-space optical communications, especially, ones employing indoor optical wireless LANs, has been developed. The receiver enables a compact and intelligent optical communication system through its two functional modes : image-sensing and communication. We clarified that the diffusion of photo-generated electrons determines the minimum total transimpedance gain of the amplifier and the maximum intensity of optical communication signals. A 50x50-pixel photo receiver with four independent communication-output channels was designed and fabricated with standard 0. 35 μm CMOS technology. To amplify optical data in the communication mode, it has in-pixel trans-impedance amplifiers and column-parallel main amplifiers. Experiments conducted with it showed that data rates of 50Mbps and 30Mbps were possible for wavelengths of 650nm and 830nm, respectively. Transimpedance gain varied from 660kΩ to 4.9MΩ for the data rate of 10Mbps. We also successfully captured images at a frame rate of 50Hz.
Organs within the body have the ability to regulate their own blood supply. Vasoconstriction decreases blood flow, while vasodilatation increases blood flow. Visualization of vasoconstriction and vasodilatation of coronary arteries with vasoactive agents is a useful research tool for evaluation of coronary vascular response. Microangiography with a spatial resolution in the micrometer range was carried out in isolated perfused rat hearts using a high-speed imaging system and a synchrotron radiation source at SPring-8. The imaging system was composed of an X-ray direct-conversion-type detector incorporating an X-ray SATICON pickup tube and a high-speed X-ray shutter. Microangiographic images were obtained and stored in a frame memory system with a 1024 × 1024-pixel, 10-bit format. In the imaging experiments, the isolated rat heart was mounted on a steel tube and perfused with oxygenated perfusion fluid. After iodine contrast agent injection into the ascending aorta, vasodilatation in small arteries was visualized in response to vasoactive agents.
We demonstrated a CMOS image sensor using pulse frequency modulation (PFM) for retinal prosthesis. We are fabricated the 32x32 pixel PFM image sensor with standard 0.6μm CMOS technology. We designed this chip for use in in vitro electro-physiology experiments. The image sensor pixel was composed of a PFM-based photosensor with a frequency limiter, a pulse shaper/amplifier, and a stimulus electrode. We measured the image sensor response vs. the illumination level. We further verified the pulse shaper/amplifier characteristics. By using the current output from the stimulus electrode, we successfully demonstrated positive and negative current output of 100μA driving an external electrode placed inside a multi-electrode array immersed in saline. This result shows that this chip could be used in vitro electro-physiology experiments. To observe the output pulse from the sensor, we used the LED display in order to visualize the input scene from each pixel output pulses, and demon-strates the visualization successfully.
This ultrahigh-sensitivity HDTV color camcorder uses a proximity-type image intensifier employing 3rd-generation GaAsP photocathodes. In order to accomodate a video cassette recorder (VCR) within the camera, thereby reduce the weight and increase the image quality, a compact optical unit consisting of an enlarger lens and tapered fiber optics plate (FOP) was developed. The camcorder can be equipped with a high-speed gate shutter with speeds up to one per microsecond by pulsing the voltage to the photocathode. The camcorder provides good color reproducibility under moonlight, and a signal level of 100% was obtained at 0.2 lx, f/2 (signal-to-noise ratio of 56dB). Its sensitivity is about 400 times greater than that of current HDTV CCD camcorders, mak-ing it particularly well suited for shooting faint objects such as auroras and nocturnal animals.
This experimental 60 frames-per-sec progressive-scan HDTV camera has 1, 920 × 1, 080 pixels and three CCDs for RGB color. The horizontal CCD of the camera was driven at 148.5 MHz, and the transfer efficiency of the CCD cell was about 99.98%. A 2/3-inch 2.2-million pixel FIT (Frame Interline Transfer) CCD for an interlaced HDTV camera was used for the progressive scanning. This involved using multiple FIT drives and storing the odd numbered picture lines in the field memory and the even numbered picture lines in the vertical signal transfer CCD. The horizontal and vertical resolutions of this camera were each about 1000 TV lines (TVLs), and the vertical MTF (Modulation Transfer Function) response was about 57% on 700 TVLs and 30% on 1000 TVLs. A computer simulated conversion from a progressive to an interlaced image that used images acquired by this experimental camera reproduced an alias-suppressed interlaced image.
Electro holography is a technique that can reconstruct moving color 3D images. Recently, several methods using a white light source or laser waves have been tested. However, since the reconstructing waves are applied closely approximated to the plane waves, the system tends to be very complicated and, hence, difficult to make smaller. In this paper, we examine a virtual image observing process using LED as a reference beam and analyze the color characteristics of the process. The results showed excellent color characteristics.
An experimental ultrahigh-definition color video camera system with 7, 680 (H) x 4, 320 (V) pixels has been developed using four 8-million-pixel CCDs. Each 8-million-pixel CCD with a progressive scanning rate of 60 frames per second has 4, 046 (H) x 2, 048 (V) effective imaging pixels, each of which is 8.4 micron2. We applied the four-imager pickup method to increase the camera's resolution. This involves attaching the CCDs to a special color-separation prism. Two CCDs are used for the green image, and the other two are used for the red and blue images. The spatial image sampling pattern of these CCDs relative to the optical image is equivalent to one with 32 million pixels in the Bayer pattern color filter. The prototype camera attains a limiting resolution of more than 2, 700 TV lines both horizontally and vertically, which is higher than that of an 8-million-pixel CCD. The sensitivity of the camera is 2, 000 lux, F 2.8 at approx. 50 dB of dark-noise level on the HDTV format. This camera system also has a new contour compensation circuit. It suppresses the false color caused by the non-linearity and the pixel-count difference between green and red (or blue). The technique has contributed to the reduction of the camera's signal process circuitry.
We propose a construction of intelligent image sensors that perform associative memory tasks with nonmonotonic CDMA neural networks. The CDMA approach allows implementing Hopfield's associative neural network on a 2-D rectangular grid (chip), while the storage capacity of the chip is increased by employing a nonmonotonic transfer function. The numerical results of the nonmonotonic CDMA neural networks indicated that i) the CDMA network performed the associative memory task well although the multiple-valued outputs were muxed and demuxed by the CDMA, ii) the number of neurons that could be implemented on a 1 cm×1 cm chip was 340 for a 0.6-μm CMOS process, and iii) the storage capacity (the number of stored patterns per number of neurons) was approximately 0.3, which implies a possible development of associative memory systems on image sensors.
The multi-projection display system “D-vision” has a hybrid curved screen for immersive virtual environments. The screen consists of differently shaped surface elements that are partial surfaces of such objects as planes, spheres, cylinders and tori. High-quality stereoscopic images are generated with 24 PCs and 24 projectors. Our rendering algorithm using image processing hardware and 3-D graphics hardware realizes a seamless and non-distorted projection on the hybrid screen. Distortion of the images observed by two or more users is also reduced by the characteristic seamless shape of the screen. In this paper, we give an overview of D-vision and the technologies used for the rendering and projection hardware. We also show the effectiveness of our system by comparing it with conventional immersive projection displays.
Designers, who develop moving images, ordinarily create three-dimensional computer animation (3-DCA) in collaboration with operators, who work using computers. In this case, the operators need to be able to receive each image motion from the designers accurately. In this paper, we propose a new work flow system for creating 3-DCA to communicate the timing information of the image motions. We present a numerical definition of the timing information for 3-DCA and propose a tool for extracting it from the designer's hand motions. Operators can then receive the numerical timing information from designers using the tool with the proposed work flow. The tool is simply composed of a personal computer and a CCD-camera. Experimental results showed that the extracted numerical timing information was similar to that of the image of the designers. The proposed work flow and the tool are useful for efficiently communicating timing information and developing 3-DCA.
Separating the diffuse and specular components from the reflection of a 3D object surface is very important for generating highly realistic synthesized images in computer graphics applications. Most researchon this separation has only considered cases in which an object is illuminated by a single point light source. However, when an object is only illuminated by one source, acquiring the correct texture information of the visible regions on the object surface can be difficult or even impossible because of shading. To overcome this problem, we used two point light sources to illuminate the object to reduce the shaded regions. We used an iterative algorithm to estimate the intensity variation curve parameters and to compute the separation of the two components. To improve the separation quality, we used a weighting function to reduce the effect of noise contained in the highlight regions. Experiments using both synthesized and real data demonstrated that our method is effective and produces clear separation.