We have developed a real-time watermarking system for High-Definition TV (HDTV) using a VLIW-based media processor. The system preserves picture quality of each video frame by using motion estimation and contour detection that optimize the watermark strength. For real-time embedding operation, the system divides an HDTV video frame into seven segments so that the independent seven watermarking modules, each of those comprises three "MAPCA" media processors, works in parallel.
When image processing techniques are used as an image inspection system of utility apparatus, they require good image quality to work well. The fog and gas, something like that, go down PSNRR of apparatuses images about 15dB. This affection is equivalent to noise added. In order to reduce noise, edge preservation noise reduction scheme has been developed. When object tracking method was applied to galloping images whose PSNRR was 11dB, a normal noise reduction scheme failed to track a target object. But the proposed scheme could work well..
Utilizing the combination of the IR (InfraRed) -tag and wireless LAN (Local Area Network), this paper proposes a new location awareness system that automatically detects and tracks the location of the mobile PC users so that each user can derive the information based on the place where he/she is. The proposed system automatically starts to aware the user's position when the user enters the system region, using DHCP (Dynamic Host Configuration Protocol) service. Each user's IR-tag is controlled by the system through the network connection and tracked IR-tag is kept associated with the network information to identify the user.
The authors study about a method for detecting television transmission failure by use of invisible markers signals. The proposed system is based on extraction of feature information of detected invisible markers and statistical test. The proposed method defines the transition of marker detection result between adjacent frames as feature information and conducts statistical hypothesis test to detect local impairment in a frame. Computer simulations show that the proposed method achieved better performance than the conventional method.
We propose a novel method for contour transformation using motion estimation of object fringe. In the proposed method, first, track points are set on the object fringe, and control points are set on the object contour. The control points are moved according to the result of motion estimation of track points. The motions of track points are estimated by the weighted least-squares method, and the reliability of the result of motion estimation is used. In order to demonstrate the efficiency of the proposed method, we applied it to object extraction from video. The proposed method reduced error of object extraction and extended tracking time. It was confirmed that this method extracts non-rigid body object in the case of complex background video.
Two methods for improvement of image-quality in JPEG2000 coding with ROI (Regions of Interest) are proposed in this paper. In JPEG2000, there is a coding unit refered to code-block. We show that it is a problem that the relationship between this code-block and the ROI region are influential in image-quality. Then we propose two methods to improve the problem. One is a method which specify the ROI region with consideration to the code-block size, the other is a method which selects the frequency-band in the DWT domain to be applied to ROI. The both proposed methods not only improve the deterioration of image-quality, but also they are capable of decoding with the standard JP2 Part-1 decoder. Finally, we provide some simulations to verify the effectiveness of the proposed methods.
In this paper, we evaluated JPEG2000 error resilience tools over an orthogonal frequency division multiplexing (OFDM) channel. The error resilience technique adopted in JPEG2000 is based on the insertion of marker codes in the codestream. It does not provide enough error protection for critical data (such as the header and data in upper layers) in a high-error-rate channel. Therefore, we also propose a new method to combat channel errors. The method exploits the layer structure of the JPEG2000 codestream. Prior to transmission, data in the top layer is duplicated and protected by a chosen forward error correction (FEC) code. This data is then embedded into the bottom layer. In the receiver, a post-transmission process is applied to the codestream that extracts the embedded data for error recovery purposes. Our method provides multilevel error protection for the top layer and preserves the same codestream structure as the one in the JPEG2000 standard. In addition, the proposed scheme can be combined with the existing error resilience technique.