Please find below some links that we hope to provide a better insight
into the R&D projects we are working on. The demonstration here is
meant only for non-commercial showcase of our technologies.
WavPRESS -- is our core wavelet-based scalable image and video compression engine. Here, you can see a contrasting improvement of WavPRESS over JPEG standard. As we will further demonstrate below, it does not only give good compression performance but also highly scalable compressed images and videos.
Image Archiving -- here, we attempt to showcase the strength of progressive browsing
(bit rate scalability). You will see a library of images. Once you have chosen a particular thumbnail image,
the full-size, high-quality image will then be streamed to you from the image server.
We also explain the various other scalability features that WavPRESS can support.
Scalability has become a timely and strategic enabling technology for many
Internet applications, especially in the present highly heterogeneous environment
(either in network bandwidth, display devices, hardware complexity, etc.)
Partial Image Retrieval -- here, you can navigate and zoom to view a small
segment of a very large image, such as a world atlas or high- resolution medical
image. [To view this interactive demonstration, you have to temporary disable
some security settings in your Internet Explorer browser as we have yet to apply
for a proper commercial digital certificate for the ActiveX plugin. To do so,
click on Tools->Internet Options on your browser. Click on "Security" tab,
and click on the globe image "Internet" and then scroll the Security Level for
this zone to LOW. Click APPLY and then OK. It is fine if you receive prompts that
the plugin is insecure, etc., as I can guarantee you that it will be safe; just that
we did not pay commercial firms like VeriSign to certify our research program.
After this, you can go back and select either the cheetah or medical image.
You can then select a portion of the large image, and the image server will extract
and stream to you only the selected region of interest.
WAFIS -- this is a wavelet-enhanced automatic fingerprint identification system.
WAFIS can identify/authenticate a fingerprint from a large database accurately and fast.
A one-to-one matching will take only about 0.3 second, and searching to identify the right
print from a database of 100 prints will take less than 1 second. These speed results are
software-only solutions. As you can see, our technology is highly robust; the prints with blur
or a large portion of the core segment missing will still result in accurate matching result.
Most commercial products would have failed to recognize the right print or reject to accept
the authorized person. Two patents have been filed on this technology. We are currently
working to build this technology into a single chip solution to improve the speed performance
and its readiness to engage commercial applications.
Industrial Fault Diagnosis
(ICMFD) -- this project focuses on a powerful set
of machine diagnosis/analysis tools using wavelet technology.
From the graphical illustrations here, you can see how well our
technology performs in comparison with other conventional/traditional
techniques. We are also keen to explore further into using this technology
in other areas such as partial discharge location for failure detection of cables
and feature extraction of biomedical signals.
Wavelet-based document image compression
(WDIC) -- this project focuses on high-quality,
high-compression storage of digital documents. As you can see, our technology,
WDIC, is far superior to JPEG and SPIHT (which is among one of the world's best
wavelet-based compression techniques). WDIC is specially designed to perform
exceptionally well with digital documents that contain both text, background, and
images. The images there clearly shows its potential for document archiving, digital
faxing from PC, and many more.
WAVC -- this project focuses a powerful set of video compression algorithms.
Combining advanced wavelet technology with the most successful motion
estimation/compensation technology, such as TRUE overlapped motion
estimation and overlapped motion compensation; this set of algorithms
provides hi-fidelity, high-compression ratio storage of digital video.
It can be used for many multimedia applications, such as video-on-demand
and internet video streaming application.