Charge transfer efficiency(CTE) in CCD image sensor is analyzed by a three dimensional device simulator SPECTRA. The limitation of CTE estimation by minimum fringing field and structure dependence of CTE are discussed. It is found that CTE of long and narrow channel CCD strongly depends on decay time constant of transfer pulse.

View full abstract

A new structure for an output circuit of HDTV CCD imager has been developed. This structure has overlapped two poly-silicon electrodes. The highest applicable drain voltage for the 0.7 μm effective channel length structure reaches to 12 V with no degradation of mutual conductance. The usage of a low concentration n layer under the drain-side electrode ensures longer than 10 years lifetime. The calculated power of an output circuit using the structure is 29 mW with 110 MHz cut-off frequency and 0.57 voltage gain.

View full abstract

A 1/4-inch format 270k pixel interline transfer(IT) CCD image sensor has been developed, which is suitable for compact size video cameras. High saturation signal of 600mv, low smear level of 0.01%, and high sensitivity of 40mv/lx have been realized.

View full abstract

A high-sensitivity 1/4" 250k pixel Interline Transfer CCD (IT-CCD) image sensor has been developed. To achieve high-sensitivity, high-energy ion implantation process is adopted for photodiode (PD) formation. According to device simulation results, transfer gate (TG) and channel stop (CS) length are reduced to increase saturation signal. Sensitivity improvement of 35% are obtained, as well as 100mV saturation signal. Low smear level of -95dB is also achieved.

View full abstract

A 1/4-inch format 320k pixel Interline Transfer CCD image sensor has been developed. High saturation signal of 1200mV and high sensitivity of 50mV/lx are obtained.

View full abstract

A two-level image sensor with very low image lag and very small fixed pattern noise has been realized. Charge ingection into the storage diode and the a-Si layer reduced image lag. A low bias voltage for the storage diode resulted in very small fixed pattern noise.

View full abstract

A dynamic electronic shutter for STACK CCD imager was newly-developed, which employs chage summation of small signal without shutter operation and large signal with shutter operation. The proposed method results dynamic electronic shutter while without decreasing sensitivity.

View full abstract

The high light signal suppression operation (knee operation) is studied in order to apply to the STACK-CCD imager. The measured photo-conversion characteristics around the knee point is good agreement with the calculated values obtained from the thermal emission current of the readout gate. The knee point uniformity of the STACK-CCD imager is superior to that of the conventional IT-CCD imager.

View full abstract

We have developed a multi-driving circuit of AMI. It can drive an AMI for interlace scanning and progressive scanning when we control only one switch, because AMI equips different elements for each function. We can easily get two different images with one AMI and same driving circuit and confirmed that progressive scanning output is higher than the other. Also the circuit has shutter function.

View full abstract

A new CCD timing generator circuits to suppress fixed pattern noise caused by digital circuits has been developed. The unique features of this circuits are to halt a timing generator during active scanning interval, and to employ a 1H-glass delay line to generate horizontal synchronizing signals. We described the effect of this circuits implemented in an HDTV CCD pickup system.

View full abstract

We have examined a new vertical pixel shift method, using two interlace structured CCDs. This method may cause a line flicer due to the diffrent integration operation of the two CCDs. In this paper, an experimental results to reduce flicker by using an optical low pass filter are described.

View full abstract

New image acquisition system for HDTV is proposed. This system, named "Quad CCD imaging system", is equipped four 2/3 inch CCD sensors, two for green, one for red, and the rest for blue. To realize it, we have developed a new dichroic prism nearly as small as the conventional one for the RGB system. It is mentioned that this system is less vulnerable to the negative effect of the chromatic aberration on the resolution than the RGB system.

View full abstract

This paper describes the features and structures of HDTV color camera which contains high performance FIT-CCDs and compact analog image enhancer. This camera is able to operate in stand alone system without camera control unit.

View full abstract

We developed a new vertical false-color-signal reduction method for single CCD color video cameras. We have confirmed by computer simulation that this method can reduce the false-color-signal without optical low-pass filter.

View full abstract

Contrast edge detection was realized by using CCDs and NOR circuits. The CCD with two input gates was used as a differential input device. The NOR circuit was fabricated based on ν MOS Tr technology.

View full abstract

The intelligent image sensor to detect an edge and a motion are strongly demanded, because of high speed and high performance processing. We have investigated the algorithms of an edge detection and a motion detection, closely related to the parallel processing. The performance of the edge detection of developed circuit is experimentally demonstrated for stationary and moving step-edge-inputs.

View full abstract

We propose a new method to generate hologram for 3-D displays. Since the spatial frequency spectrum of a piece of the hologram has depth-added 2-D image of recorded object, we can generate the hologram by transform from the spectrum data which is easily obtained from 3-D object. Experimental results are also shown.

View full abstract

It is not practical to apply 2-D image coding to the holographic fringe because the character of the fringe is quite different from 2-D image. We demonstrate 2-D coding can be used effectively to compress fringe when the segmented fringe is transformed from spatial region to frequency region.

View full abstract

One of the special feature of the kinoform using the LCD panels is that color moving image can be reconstructed with plural LCD panels. This time at first we construct the kinoform using the input pattern resolved into colors and display the kinoform to each LCD panels. Then we make clear that we can reconstruct the color image at same plane with lasers of different wavelength and we consider the characteristics of the reconstructed image.

View full abstract

We have developed a videodisc system which reproduces Hi-Vision signal without bandwidth compression, and 4-channel high quality digital audio. Using 3-beam recording and reproducing technics, the maximum playback time of 60 minutes was obtained with both sides of the disc.

View full abstract

We have developed a videodisc player which playbacks Hi-vision baseband signal with even-powered 3-beam optical readout. To readout playback signals, all beams have to be on the tracks, while no tracking error signal can be obtained. We solved this problem by shifting the side beams by 0.1 μm to the center track.

View full abstract

The master disc recording machine with three recording beams has been developed. This machine is able to record three different signals at the same time using three beams. Consequently, high baud rate data such as Hi-Vision Baseband signals are able to be recorded on an optical disc.

View full abstract

An optical head using 680nm laser diode has been developed for magneto-optical disk systems. Using 680nm laser diode, we obtained 1.01μm x 1.38μm beam spot. Read/write characteristics in 1.2μm track pitch disk were examined. In a result, 46dB CNR was achieved at 0.91μm pit length.

View full abstract

In response to the recent demand for high density recording in an optical disc, "going narrow track" is now in progress. We have developed an adaptive learning control method capable of changing its learning ability in accordance with track correlation, with which we can avoid unnecessary learning during the vibration from the equipment and also the lack of stability which is the problem for learning control, and furthermore we can increase follow-up ability. As a result we have confirmed that track-error could be minimized almost to a "0" level and stabilized operation with repetitive learning control system could be achieved.

View full abstract

1,7 codes are not sensitive recording codes to intersymbol interference. It is possible to be rid of the nonlinear distortion by the 1,7 codes + RAMDFE+Viterbi Equalizer. Performance of this EQL has been confirmed by our simulation before. We developed this EQL system with one ECL G/A.

View full abstract

In order to develop the compact Hi-Vision digital VTR for field program production, the bit rate reduction technique is indispensable. We have been studying the degradation of picture quality caused by the iteration of picture coding, whose method is intra-field DCT based on JPEG, with the signal processing such as picture shift or gain change. As a result of the simulation, S/N was degraded according to the coding generation and the degradations varied with the shift size or gain.

View full abstract

A new PRML system with optimum intersymbol interference sequence (ISI) are described. Firstly, an error probability function of PRML system with an arbitrary ISI was derived. Secondly, the optimization of ISI conducted so that error probability calculated with the function is minimized. The PRML system with the optimized ISI has higer performance of over 4.8dB than conventional PRML systems.

View full abstract

This paper reviews digital signal error counter for digital video cassette recorder. Conventional error rate measurement system supported one of various 8-n modulations and only playback mode. We have developed error rate measurement system for various 8-n modulations and speed search mode.

View full abstract

To attain a high recording density, we studied application of APSK (Amplitude Phase Shift Keying) for magnetic recording channel. New mapping (7,1) APSK that has optimum allocation of recording signal points for magnetic recording is developed. By employing (7,1) APSK, a linear packing density as high as 129 kbpi (bit length of 0.2 μm) is accomplished with MP tape.

View full abstract

NEC has developed Hi-Vision Digital VCR for consumer use. The feature of this VCR are 2-houres record and playback using A6 size caseete. This paper describes outline of DRUM for Hi-Vision Digital VCR.

View full abstract

This paper is about a study of tape/head contact on VCR. We established measuring system by transparent tape, and applied it to measuring tape/head contact with high scanning speed as 50m/s of a head. The relation between tape/head contact and envelope of output level is defined.

View full abstract

In order to analyze the causes of the reproduced output variation on VTRs induced by spacing loss and tracking error, we studied a measuring method which utilizes a swiching signals between two different frequencies. As a result, this method has proved useful for obtaining the quantitative assessment of output variation causal factor.

View full abstract

To evaluate VCR drum assemblies, the tape-to-head interface evaluation system has been developed. With this system, the thickness of the air film over the entire wrapping area of the drum can be measured. Based on experimental results, drum design has been optimized, and the accuracy of the calculated value of air film thickness has been confirmed.

View full abstract

We have developed the noise-less continuous slow playback method for time lapse VCR. We introduce a new playback method to realize a noise-less reproduced picture in the continuous slow mode by azimuth predict change according to the track trace of magnetic heads.

View full abstract

A new variable speed playback system with dynamic track following was developed, using two actuators with double azimuth heads, head-sellector by which head scan under new sequence and V-dancing compensator. In this system, a field/frame playback with high picture quality and smooth motion at near normal speed was realized.

View full abstract

An efficient two-pass coding scheme has been developed in photo transmission system. In this scheme, optimum bit assignment using downsampled image is adopted to source image, so that each image is always coded by optimum quantization table. In this paper, the bit assignment procedure is described.

View full abstract

In three color component images (i.e.R, G, B), the separate components have a strong correlation with each other in small areas. Using Principle Component Analysis, three new color coordinates can be built. The data of these newly built color components is such that the principle component of the three contains an extremely high concentration of the image and is not reliant on the other two components. Therefore, the two components can be eliminated without degradation of picture quality. Quantization of transformation coefficients is discussed.

View full abstract

The previously developed universal pattern-matching coding technique(PCM) uses different coding modes, and adaptively chooses the proper coding mode instantaneously for a given input subblock. To improve the adaptability of the coding mode selection, we introduced the delayed-decision mechanism with the (M, L) algorithm. The work presented here optimizes the values of its search parameters from the standpoint of the coding performance.

View full abstract

We have developed a new method to improve the accuracy of the quantity output DCT-based (JPEG) image data compression using Fuzzy Logic Algorithm(FLA). Several rules are prepared to calculate the gain factor that drives the quantization table. The average accuracy of the amount data ouput coder from the simulation result is -1.5% (the error range from -3% to 1%), without FLA the error range from -41% to 43%.

View full abstract

ABS method algorism of a constant coded data, is developed based on adaptive block size coding. An experimental hardware has been implemented that is capable of real time processing and recording of High Difinition pictures (i.e 1344 x 960 pixel) to D-2VTR. The S/N ratio of the decoded picture shows more than 40dB at the compression ratio of 1/10.5.

View full abstract

In the KL-transform image coding scheme based on variable block shapes, basis-vectors are individually derived for each block from the isotropic autocorrelation function model. This paper proposes a new method to select adaptively an autocorrelation function from among the isotropic model and four kinds of directional ones. Simulation results indicate that this scheme can provide 0.5〜1.0dB additional improvement in SNR over the above nonadaptive scheme.

View full abstract

The KL-transform image coding scheme based on variable block shapes can diminish remarkably the mosquito noise, because it can fit well block boundaries to the principal contours in an image. To do so much better and to improve both SNR and reproduced image quality, in this paper, movable areas for vertices of each block are magnified.

View full abstract

This paper proposes a new region-oriented coding scheme employing orthogonal transforms. First, an image is partitioned into about 100 regions, some of which are redivided into square-blocks and variable-shape-blocks. Then, a mean value and the AC component in each block are encoded by DPCM and OTC, respectively. This scheme can achieve better coding performance than the JPEG baseline system.

View full abstract

Subband filters satisfying both orthogonality and phase-linearity are generally derived for L=2N, where L is the filter length and N is the number of channeles. These filters include the LOT and the maximum coding gain of the subband scheme with the filters proposed here exceeds that of the LOT.

View full abstract

The subband scheme employing the emphasis and ideal band-splitting to obtain the maximum coding gain is approximated and the coding simulation is performed for super high-definition images. It is shown that the emphasis is very effective to increase the SN ratio and reconstruct the edges of the images well.

View full abstract

Subband filter characteristics are evaluated by the subjective test of reconstructed pictures in subband coding shemes. It is concluded that the subband employing QMF is superior to the DCT, linear-phase filters are superior to the nonlinear-phase filters and short length filters are superior to long length filters.

View full abstract

We are researching Cooperative Working Support System operating together with Multipoint Teleconference Systems and computers. We developed Cooperative Working Support DRIVER(CWSD) by which muktiple terminals in a teleconference can use the same application program.

View full abstract

Conversion from input texts to moving facial images and speech has gained attention as a basic technique to realize new types of communication services and human interface. This paper focuses on the control method for image synthesis.The proposed control method can accept multiple input sentences and more control words, and thus can synthesize more various and longer facial image sequences easily and automatically.

View full abstract

Recently, there are some studies of applying motion compensation to the moving region which is detected in order to compensate the motion of an object more correctly for moving picture coding at low bit rates. We have developed a moving region detection algorithm which searches the edges of moving region from the top and the bottom and from the left and the right according to adaptive threshold processing, and defines the inner side as the moving region.

View full abstract

It is well known that the 3-D TV system requires a tremendous amount of data to generate the stereoscopic visual effect. To date most of the work on the 3-D image data compression concentrates on stereo images. In this paper, we show the experimental results of the Multi-View 3-D image data compression, which employs the concept of disparity compensation and is based on the standardized coding scheme H.261.

View full abstract