ADDING DEPTH TO EN­TER­TAIN­MENT

YOU ARE ABLE TO WATCH 3D MOVIES ON YOUR TV AT HOME. BUT HAVE YOU EVER WON­DERED HOW IT WORKS?

Gadgets and Gizmos (India) - - KNOW TECH - TEXT AND GRAPH­ICS BY SAN­TOSH KUSH­WAHA

To see stereo­scopic 3D im­ages, the left and the right eye need to see dif­fer­ent im­ages. How­ever, even with­out this ef­fect it is pos­si­ble to cre­ate the il­lu­sion of depth in im­ages. With their sheer size, screens in IMAX the­aters con­vey a sense of per­spec­tive that feels three di­men­sional. And be­cause we are all so ac­cus­tomed to watch­ing tele­vi­sion, our brains per­ceive the 2D im­ages we see there as three-di­men­sional to some de­gree. You will need to buy a 3D Blu-Ray Disc player along with the 3D TV to view con­tent. 3D-ca­pa­ble TVs cost twice as much as reg­u­lar Full HD panels of sim­i­lar sizes. The gen­eral prin­ci­ple be­hind IMAX’s 3D-like im­ages also ap­plies to the new 2D-to-3D con­ver­sion func­tion on some TVs. The idea is to cre­ate a sharp con­trast on screen—be­tween a large and small ob­ject, through over­lap­ping ob­jects, or through dif­fer­ing lev­els of bright­ness or clar­ity—which then forces the right and left eye to an­a­lyse things dif­fer­ently.

An­other pop­u­lar 2D to 3D con­ver­sion is ImageIQ3D from HDLogix us­ing so­phis­ti­cated mo­tion anal­y­sis called op­ti­cal flow. Much in­for­ma­tion about the 3D scene ge­om­e­try can be gleaned from the rel­a­tive mo­tion of ob­jects in the video and how they oc­clude (re­veal and hide) pix­els in other ob­jects as they move—as long as the mo­tion es

tima­tion is pre­cise and ac­cu­rate. Sec­ond, straight lines of build­ings, the hori­zon, and other ob­jects gives clues about the van­ish­ing points, which also help to solve the puz­zle. A gen­er­alised hough/radon trans­form helps iden­tify th­ese lines and other use­ful fea­tures. Fi­nally, in most pho­tog­ra­phy there is a ten­dency for ob­jects that are very near and very far from the cam­era’s fo­cal plane to be blurred by an amount pro­por­tional to their dis­tance from it. A Blind Point-Spread-Func­tion Es­ti­ma­tor is used to es­ti­mate the out-of­fo­cus char­ac­ter for each pixel to com­plete the in­for­ma­tion needed to es­ti­mate the depth of the video. Some of this in­for­ma­tion is al­ways avail­able, some­times not all of it is (for ex­am­ple, when noth­ing is mov­ing in the video). ImageIQ3D uses a su­per­res­o­lu­tion-based sta­tis­ti­cal ap­proach to achieve ro­bust and con­sis­tent re­sults even when there is very lit­tle or par­tial in­for­ma­tion avail­able. Ul­ti­mately, the goal is to pro­duce an ac­cu­rate depth map for each video frame—a rep­re­sen­ta­tion of the dis­tance of each pixel in the video from the cam­era. Once an ac­cu­rate depth map is cal­cu­lated, it is pos­si­ble to eas­ily con­vert to and from

any 2D or 3D for­mat.

Shift­ing Di­men­sions PER­SPEC­TIVE BASED ON SIZE PER­SPEC­TIVE BASED ON OVER­LAP PER­SPEC­TIVE BASED ON BRIGHT­NESS PER­SPEC­TIVE BASED ON CLAR­ITY

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.