Espier

Icon

No theme, Just views

What is IPad? Its the New Newton!

What is the difference?

Affectionate

Youngsters has grown in the expressing their love to their partners. Look what I found under the signboard somewhere near Serdang. It says:
Birthday Banner

CR,
Happy Birthday, I love you
Ken.

Anyone know CR / Ken, please comment here! I would like to meet this couple!

Green

While I was searching for suitable photo to used in my upcoming paper. I found this in my Picasa. A lovely photo taken by my dear with my LX3. Click to enjoy.

P1020327.JPG

Compressed Sensing

Compressed Sensing has been actively research in the last 5 years since 2004 when it was introduced by Donoho, Candès, Romberg and Tao. The motivation behind compressed sensing is information can be compressed when it is being sensed and can be reconstructed with minimal data. In short, more data can be obtained with this technique hence increase the wealth in information.

Idea

The idea behind compressed sensing is from the fact that the sparsity of useful information during data collection or sensing. From the data sparsity, a mathematical calculation can be used to calculate and reconstruct the data into useful information from the linear form of data sparsity.

Let me put them in layman view. Say you have a digital camera to capture images. The digital camera obtain the data from its sensor in RAW format (which means every bit of light is capture without any compression) and then only compressed to JPEG (conversion technique that convert RAW to a smaller size of image) to be stored into your storage. Typically, the size of a 10 megapixels RAW image is around 10+MB. On the other hand, the JPEG image of its equivalent is about 1/5 of its uncompressed brother. In this scenario, we actually wanted the 4MB information but we are capturing them at 10MB. The 6MB wastage of data includes noise, redundant data…

Hence, compressed sensing is about sensing them in compressed form right from the start and using high order mathematical formula to reconstruct them to the information we wanted. Theoretically, the performance of compressed sensing can be as good as reconstruct a data of only 10% of its original counterpart and reconstruct them back to full scale 100% original signal. Impressive isn’t it? This might not be far away, they will be with us in 5 to 10 years.

Read the rest of this entry »

Hailam Kopitiam

image

My first shitbrix post are brought to you with my Milestone.. look at the chicken rice.. the chicken is dry and dark..

Links