The improvements in the image output characteristics of an image sensor with a function for detecting a modulated light signal is described. Three effects that degrade the characteristics were identified by examining the output characteristics in detail : “parasitic N-diff effect”, “diffusion carrier effect” and “residual charge effect”. Two kinds of improvements in modifying operation timing and employing in transverse overflow drain structure in a pixel were introduced in a new version of a CMOS image sensor based on the examination. An image sensor with 64×64 pixels based on the improvements was fabricated using 0.6μm CMOS technology. The gray scale characteristics of the sensor improved during a demonstration.
A new method for producing interpolated signals between a sampled green signal of a single-imaging device color camera with a Bayer array of primary color filters was developed. The method is accomplished by extracting a block of 8-by-8 signals so that sampled green signals exist on all the pixel locations of the block by rotating the axes of the block. The method provides a set of frequency coefficients using a discrete cosine transformation with the block. It also provides interpolated signals that correspond to the interpolation location using inverse discrete cosine transformations with the set of frequency coefficients. The method can reduce errors in the signals by more than 18dB compared with the conventional linear interpolation method.
In this paper, first, we describe the improved phase-correlation method used to measure the amount of horizontal shift between the left and right images of stereoscopic HDTV programs. Second, we measured the amount of horizontal shift of 48 stereoscopic HDTV images using the method. Third, we carried out a subjective evaluation test to assess how comfortable (or uncomfortable) each of these images was to view. The results indicated that the comfortable viewing range of parallax distribution in a frame was within approximately 0.3 diopters.
A compact image capturing system called TOMBO (thin observation module by bound optics) is proposed. This system is an opto-electronic hybrid system, that utilizes a compound-eye imaging optics to achieve a very thin system configuration. Digital processing of the captured multiple images is used to retrieve the object image. Experimental results using the TOMBO prototype system are presented. As an application, a compact fingerprint capturing system using the TOMBO is described.
We have been investigating an integration of sensing and compression in an image sensor. A compression Sensor reduces the number of pixels in the image signal that have to be read out from the sensor, We present a real-time object tracking system by using a compression sensor that has 128×128 pixels and FPGA, By using this system. several moving objects can be extracted and tracked independently at 1200 frames/second. We also describe real-time depth estimation by using a binocular compression sensor. We show some results obtained by the system.
Copyright protection of broadcast content has become an issue of great concern because digital broadcasting is increasingly used to deliver high quality visual and audio content. We analyze the existing copyright protection technologies and clarify why they cannot prevent illegal copies from being created through the analogue output of broadcast receivers. Illegally copied content can thus be re-digitized and distributed through the Internet and portable media. Digital watermarking that survives analogue processes effectively complements existing technologies. Four methods of using watermarking to detect illegal copies, copiers, and distributors are described and compared with respect to their implementation cost, detection ability, and ease of deployment. The technical requirements of the watermarking used in these methods are also clarified with respect to survivability against analogue processes, maintenance of image quality, and quantity of embedded information.
An access control system that automatically logs in a user by his or her facial information is described. Furthermore, a method for automatically tracking logged data, based on a repetitive operation of template improvements, is shown. Face recognition is a favorable variety of biometrics for personal identification because user's facial images can be captured successively from a standard video camera at a distance. Two models of the system using face recognition were tested to prevent fraud and to create automatic registration. Experimental results for a six-month test showed that the tracking capability of the method is quite practicable; the equal error rate for false rejection and false acceptance (EER) is only about one percent.
With the popularity of cellular phones featuring built-in cameras, it is no surprise that content for information transfer using images taken from these phones is in great demand. Therefore, the need for correspondent image processing technology and a system to support content development is becoming much more urgent. A new method of image matching, which contains two approaches, is described. The first approach is to segment objects from images having a relatively simple background by applying a method for growing recursive regions and for the matching. The other is to do matching directly by using a presented GA algorithm in images with complex backgrounds to extract remarkable characters such as magazine titles and brand names on commodities, which are difficult to be segmented from their environment using existing methods. Experimental results show that the approaches work effectively in our system.
We studied the degree of visual fatigue in subjects watching HDTV stereoscopic images. The images used as visual stimuli contained only absolute, with no relative, parallax. The examination of visual fatigue was carried out with a five-grade subjective evaluation test and measurements of the accommodation response after watching the images for one hour. We found that when they were displayed within the corresponding range of depth of focus and remained static in the depth direction, the degree of visual fatigue was almost the same as that induced by watching the images displayed on the screen. However, visual fatigue was clearly induced, when they were displayed outside the corresponding range. Moreover, even if they were displayed within the corresponding range, visual fatigue was induced if they were moved in depth according to a step pulse function.