A Review and Comparison on Different Video Deinterlacing
Total Page：16
File Type：pdf, Size：1020Kb
International Journal of Research ISSN NO:2236-6124 A Review and Comparison on Different Video Deinterlacing Methodologies 1Boyapati Bharathidevi,2Kurangi Mary Sujana,3Ashok kumar Balijepalli 1,2,3 Asst.Professor,Universal College of Engg & Technology,Perecherla,Guntur,AP,India-522438 [email protected],[email protected],[email protected] Abstract— Video deinterlacing is a key technique in Interlaced videos are generally preferred in video broadcast digital video processing, particularly with the widespread and transmission systems as they reduce the amount of data to usage of LCD and plasma TVs. Interlacing is a widely used be broadcast. Transmission of interlaced videos was widely technique, for television broadcast and video recording, to popular in various television broadcasting systems such as double the perceived frame rate without increasing the NTSC [2], PAL [3], SECAM. Many broadcasting agencies bandwidth. But it presents annoying visual artifacts, such as made huge profits with interlaced videos. Video acquiring flickering and silhouette "serration," during the playback. systems on many occasions naturally acquire interlaced video Existing state-of-the-art deinterlacing methods either ignore and since this also proved be an efficient way, the popularity the temporal information to provide real-time performance of interlaced videos escalated. but lower visual quality, or estimate the motion for better deinterlacing but with a trade-off of higher computational cost. The question `to interlace or not to interlace' divides the TV and the PC communities. A proper answer requires a common understanding of what is possible nowadays in deinterlacing video signals. This paper outlines the most relevant methods, and provides a relative comparison. Keywords: Visual Quality, Frame rate, Video Signals, bandwidth, Interlacing, Deinterlacing. I.INTRODUCTION Interlacing is a video system dating from the 1940’s when various television systems were defined, all of which were naturally analog. The camera sweeps the uneven numbered lines first, generating the first field and then sweeps the even numbered lines to generate the second field. As a consequence each image frame is built from two fields that do not occupy the same instant of time. Various artifacts can appear in the image due to the sampling structure of interlaced systems. For Fig 1: The Basic Deinterlacing method. example, aliasing can occur in the presence of high vertical frequencies and a specific blinking known as flicker can Interlaced video however, complicates many tasks pertaining appear in the image. Figure 1 shows a graphical representation to image processing and has its advantages as well as of the sampling structure of an interlaced system. As early as disadvantages [4]. As scanning format conversion was 1935, it was well known that the human visual system is more necessary for international TV broadcasting, the first sensitive to large-area flicker than flickering detail [1]. proposals of deinterlacing were considered only for Television broadcasters in the early years of media boom used international programme exchange. With the advent of this idea to their advantage by transmitting interlaced videos modern technology such as high-definition television that reduced bandwidth usage. Interlaced videos are videos (HDTV), video streaming, and DTH the need for a stable and (that are rectangularly sampled) scanned in such a way that on a standard conversion between formats is increasing. Modern any given frame of ‘N’ rows, only ‘N/2’ alternate rows are display systems like LED and plasma displays work on scanned. The remaining rows are scanned on the next frame. progressive video. Progressive video unlike interlaced video Volume VIII, Issue V, MAY/2019 Page No:2506 International Journal of Research ISSN NO:2236-6124 contain the entire frame in the videos’ original resolution. The Bob is the simpler of the two. Each frame of interlaced video widespread usage of displays that require progressive videos has only one half the lines. For example, the odd lines (1,3,5, has made deinterlacing an important process in the video … 479) would have pixels, and the even lines (2,4,6, … 480) processing arena. With the advent of the PC era, Haan and are blank. On the following frame, the even lines have pixels, Bellers raised the question of whether to interlace or not to but the odd lines are blank. The simplest bob interlacing is to interlace [5]. While the PC community believes that the just copy the pixels from the line above for blank even lines present-day technologies are powerful enough to produce, (copy line 1 to line 2), and copy the pixels from the line below transmit and display progressive video, the television for blank odd lines (copy line 2 to line 1). Another method community believes that it is advantageous to have interlaced would be to interpolate between the two adjacent lines to fill videos in service. The main argument put forth by the PC in a blank line. Both of these methods are shown in Fig. 2. community is that interlacing a video introduces a trade-off between vertical resolution and the time resolution. In the 1997 WinHec conference the cofounder of Microsoft Corporation, Bill Gates put forth a proposal to stabilize the picture rate of PCs to 60 Hertz progressive where it stands till date [6]. In [5], the authors studied alternate options to deinterlacing. Most importantly the question of whether present day technologies are powerful enough to deinterlace satisfactorily. The support towards an all progressive system is increasing as experiments show that an all progressive systems produces at least as good an image quality as an all interlaced system does [7]. This means that an interlaced system cannot produce any increase in quality but only an increase in performance. With the reasoning that the modern day technologies can support powerful deinterlacers, a combination of interlaced and deinterlaced systems are currently in use. In such systems, for broadcasting and Figure 2. Bob verses weave deinterlacing. transmission purposes, an interlaced system in used. The receiver has an in-built deinterlacer pre-display unit that This method can cause blurring of images, because the vertical deinterlaces the interlaced video. This also satisfied the TV resolution has been effectively halved. Weave deinterlacing community by removing the requirement for a new broadcast creates a full frame from the separate interlaced frames with protocol for progressive video. However, this now applies odd and even lines. It then copies this frame twice, to achieve additional pressure on the video processing community to the 60 fps rate. This method tends to work only if there is little come up with deinterlacing systems that are not only change in the odd and even interlaced frames, meaning there is computationally efficient to support real-time video little motion in the video. As the odd and even frame pixels deinterlacing but are also powerful enough in terms of belong to different instances in time (1/60th of a second deinterlacing performance. Deinterlacing is the task of difference), rapid motion can result in jagged edges in the converting a naturally or an artificially interlaced video into a images rather than smooth lines. This is shown in Fig.3. progressive video. Simply put, deinterlacing is the task of converting fields into frames. This process of deinterlacing is shown in Figure 1-2. The implementation methodology discussed in Figure 1-2 is the general process of deinterlacing, though the process of deinterlacing depends on the type of interlacing. It is reasonable to assume from deinterlacing literature that a video is interlaced while acquisition, owing to the virtue of the data acquisition system. In this assumption, videos can be considered to be interlaced in two ways. II.DEINTERLACING An interlaced video stream is usually converted to progressive for image processing, as well as to for display on nearly all computer monitors. Deinterlacing must be viewed as interpolation, for the result is twice the video bandwidth. There are several methods available for deinterlacing, which Figure 3. Deinterlacing effects. can result in different video qualities under different circumstances. The two basic methods are known as “bob” Both of these methods have drawbacks. A better method, and “weave.” which requires more sophisticated video processing, is to use motion adaptive deinterlacing. Volume VIII, Issue V, MAY/2019 Page No:2507 International Journal of Research ISSN NO:2236-6124 Where there is motion on the image, the bob technique works The advantage of this algorithm is that it achieves the highest better, and slight blurring is not easily seen. In still areas of the resolution when there is little movement in the video. image, the weave method will result in crisper images. A However when there are large movements the field replication motion adaptive deinterlacer scans the whole image and method can generate artifacts that are highly visible. detects areas of motion, by comparing to previous frames. It will use the bob method in these areas of the frame and use the 3.2 Spatial Interpolation Deinterlacing weave method on the remaining areas of the frame. In this way, interlaced video can be converted to progressive with Another basic member of the deinterlacing algorithm family is little loss of quality. spatial interpolation, which consists of using the lines present in the field currently being displayed to estimate the missing III.LITERATURE SURVEY lines. The simplest of these algorithms is known as Line Average (LA). One descendant of Line Average that merits Deinterlacing is a well-studied topic. The last century has seen attention is an algorithm that performs a search for the borders a lot of deinterlacing algorithms being proposed. The literature of the image to perform spatial interpolation, Edge Line of deinterlacing can be studied under two broad categories, Average or ELA. These algorithms reduce the stepped effect Linear and Non-Linear deinterlacer.