We have developed thin film laminated head for high density recording. Magnetic core of this head is formed by electron-beam evaporated method on a slope which is processed V-shaped groove for non-magnetic substrate. Output at low frequency is over 4dB in comparison with ferrite head and rubbing noise is about 0 dB. This head is higher productivity than parallel type thin film laminated head by V-shaped groove processing at one block. Electron-beam evaporated technology was established and substrate materials was suitable selected. It has been achieved that C/N ratio is over 2.4dB at 1MHz and over 4.5dB at 7MHz in comparison with ferrite head.
In many digital VTR, play back signal is equalized to Nyquist's criterion, and detected to digital signal by integrated detection. We digitized play back signal after equalization with 4 times multiplied synchronous clock, and transfered these data to computer. Then we investigated the error occurring waveform and the number of errors. From this investigation, we developed new signal detection method using second difference between each bit. Threshold level is controled with the value of second difference as to increase signal to noise ratio. Detection errors near threshold level in integrated detection are extremely reduced and total error rate is decreased about 1/10.
We need much time and effort to understand video informations. The automatic extraction of useful scenes or image areas from moving image finds various and attractive video communication services. For example, by detectting the person looking at something we can count attention-rate automatioally or reserch how to look at the display by human eyes in the exhibition hall. This paper presents the detection method of visual direction from facial image. This method can detects it from low resolution facial image by real-tiem rate. The detection accuracy of this method investigated experimentally is reported.
We have developed a proto-type codec of contribution quality according to the CCIR Rec.723. The major purpose of the development is to optimize parameters which are not described in the recommendation using the hardware and many types of long time test sequences. New technology which is not included in the recommendation such as inter-field motion compensation and motion estimation using DCT coefficients can also be tested using the proto-type codec.
Transform error produced by the iteration of DCT procedure is studied. Input precision of 8 bits is assumed for the DCT procedure. DCT coefficients having a precision of 13 bits performed error free iteration. Almost the same results were obtained in the case of coefficient precision of 12 bits. Coefficient precision of 11 bits or less caused relatively big mean squared error (M.S.E.). In the case of the precision of 11 bits, M.S.E. gradually increased during iteration. On the contrary, M.S.E. showed constant values during iteration in the case of the precision of 10 bits or less.
Line type thermal printing method onto plain paper is developed. In this method, a thick-film thermal print head which has partial glaze is used. As a result, good contect between ink and paper is realized. And, a type or a property of ink is optimized. Using this method, a page printer whose performance is nearly equal to laser printer is developed. Resolution is 300 DPI. print speed is 4 pages per minute, and good printing onto plain paper is possible.
In case of the consumer use sublimate thermal dye transfer type video printer, there are additional functions such as automatic printing quality adjustment required. For example, it is not enough to adjust printing quality on several camcorders, therfore to be adjusted on the printer, again. Accordingly, in order to avoid manual adjustment of brightness, contrast, and white-balance controls on video printer, we have developed automatic exposure (AE) and automatic white-balance (AWB) adjustment functions which used Fuzzy Reasoning for picture conclusion with datas by dividing area.
Image processing techniques used in a newly developed video printer are described. The data size of chrominance signals can be reduced to one quater by using subsampling and DPCM. However, if the size is reduced to one quater by using subsampling alone, the degradation of the image quality is noticeable. Compressing luminance data to a half also brings about degradation. A field interpolation algorithm that takes into account the correlation between diagonal dots is proposed. This algorithm is simple and has better spatial frequency characteristics than the bi-linear algorithm. A color signal conversion circuit that enables users to adjust the color, tint, contrast and brightness after capturing an image is also descibed.
We have developed signal processing technology for improving video hardcopy image quality. This technology is composed of (1) adapptive pixel interporation technique, (2) automatic contrast control, and (3) saturation control for retouching
The international movement of standardization for color exchange in high density images are outlined. Since ISO/TC130 (ISO Technical Committee) reconstructed in 1989,the standardization items about various parameters and the method of measurement in the printing processes, pre-press and the material used in the process have been discussed. The standard magnetic tape format for color data exchange between devices and the high density standard color image set for print test and color analysis will be expected to come out in near future as the result of the above activities.
Recently, as the caicuiation speed and storage capacity of computer become more powerful, we can perform image processing in three dimension by computer. In the present paper, a method of analysis and recognition of knotted codes in three dimension space is proposed. The structure of the knotted codes is analyzed, and the recognition processing is displayed to show computer simulation.