EP0826196A1

EP0826196A1 - Theft-resistant video browsing using filtered versions of compressed sequences

Info

Publication number: EP0826196A1
Application number: EP96915545A
Authority: EP
Inventors: Boon-Lock Yeo; Minerva M. Yeung; Wayne Wolf; Bede Liu
Original assignee: Princeton University
Current assignee: Princeton University
Priority date: 1995-05-09
Filing date: 1996-05-07
Publication date: 1998-03-04
Also published as: CA2219789A1; WO1996036013A1; JPH11505093A

Abstract

A method and apparatus for providing theft-resistant video browsing includes storing video programmed material in a compressed motion JPEG or MPEG format (30), for permitting a user to select a video program from the stored video material for browsing (24), whereby before presentation of the selected video program to the user, the video program is reconstructed into a lowpass and/or highpass filtered version (28) thereof, to reduce the quality of the video program presented to the user for browsing to a level sufficient for permitting the user to determine whether to purchase or rent the video material, while insuring that the level of quality is insufficient for permitting the user to make a viable copy of the selected video program, or to satisfy a need to see the content of the selected video program. The video browsing system (20) and method further provides for a user to obtain progressively higher quality for the video program to be browsed by paying aggressively high fees (26) for such browsing.

Description

THEFT-RESISTANT VIDEO BROWSING USING FILTERED VERSIONS OF COMPRESSED SEQUENCES

Related Application

This application is related to co-pending application Serial No. 08/382,877, filed February 3, 1995, for "Method and Apparatus

For Video Browsing Based on Content and Structure". The teachings of the related application are incorporated by reference herein, to the extent they do not conflict herewith.

Field of the Invention

The present invention relates generally to methods and apparatus for prepurchase browsing of video material, and more particularly to methods and apparatus to limit the content of video clips without detracting from providing a useful browsing system to a prospective customer.

Background of the Invention

The ability to browse through a large amount of video material and find relevant clips is extremely important to most video applications. In interactive TV and pay-per-view systems, customers want to see sections of programs before renting them. While prepared trailers may suffice to publicize major motion pictures, episodic television, sports, and other programs will probably require browsers to let the customer find the program of interest. In the scholarly domain, digital libraries will collect and disseminate moving image material. Many scholars, such as political scientists, psychologists, historians, English scholars specializing in film, etc., study moving images as primary source material. They require browsers to help them find the material of interest to them and to analyze the material themselves. Browsing is even more important for video than for text-based libraries because certain aspects of video are hard to synopsize. Once the user has specified a query, the database may return many possible matches. At this point the user must decide which of the presented video sequences are appropriate. The patron may decide to view the video sequences in order to select the appropriate one or to revise the query.

Any browsing system must protect the associated copyrightable material or intellectual property from misappropriation by browsers or customers, while encouraging customers to select and purchase the material. The owners of the copyrightable material are unlikely to approve any browsing system to be implemented therewith unless such protection is inherent in the system.

Browsing, when used as a sales tool for video sequences, must balance the needs and wants of the customer and merchant: the merchant must offer the customer a preview to allow the customer to select what he or she wants; but the merchant does not want to give away the product without payment from the customers. While key-frame browsing discourages theft by not giving the customer the full clip in advance, it may discourage the purchase of some clips because it is not full motion video. Perceptual studies show that people can recognize important features from heavily-filtered image sequences. The present inventors discovered that such sequences can be generated directly from a compressed video sequence, saving both bandwidth and computation time.

Early browsers were developed for video production and as front ends for video databases. Today's standard technique for browsing is storyboard browsing in which the video information is condensed into meaningful snapshots representing shots while significant portions of the audio can be heard.

One known browser divides the video sequence into several equal segments and denotes the first frame of each segment as its key frame. This approach provides a rather hit-or-miss view of the contents of the video. Another known browser stacks every frame of the sequence. This browser provides the user with rich information regarding the camera and object motions. However, a scholar/researcher using a digital video library or a customer of a pay-per-view system is more likely to be more interested in the contents (who, what, where) than how the camera was used during the recording. To address the content-based browsing needs, a third known browser uses Rframes (representative frames) to organize the visual contents of the video clips. Rframes may be grouped according to various criteria to aid the user in identifying the desired material. By using the snapshots and representative frames, storyboard browsers can prevent the majority of video data from falling prey to intellectual property thefts, however, all these storyboard browsers inherently eliminate motion from the browsing copy. For most users, motion is an important aid to recognition and understanding of video. Browsing of sports, action sequences in news, and entertainment may require motion to identify the material of interest. Even scholarly collections may offer some material on a pay-per-use basis, much as the Vanderbilt Television News Archive and the Purdue C-SPAN Collection. Organizations which own the copyrights on this material may sell some of it in secondary markets and wish to maintain its value, they will often charge users small sums even for non-commercial use. Browsers may also provide playback facilities which use key frames as one method of nonlinear access to the source material. Browsing of pay-per-view material requires a scheme which encourages the customer to rent the source material but not to use the browsing version of the material. If the original source material is shown in its entirety (or even a substantial fraction of the source's length) during browsing, the customer may be tempted to view the material once during browsing without paying the rental fee; the customer can also videotape the screen during browsing to obtain a permanent record of the browsing material. If one is to provide full-motion browsing copies of video material, they must provide some way to limit the amount of information in the browsing copy. Perceptual studies show that people do not need all the visual information to recognize important features of images and that the qualities which viewers classify as high image quality do not match those features required for accurate recognition of the video material. Key frames are a form of temporal sampling of the video material.

Summary of the Invention

With the problems of the prior art in mind, the present invention provides an improved method and apparatus for pre¬ purchase browsing of video material. The customer is allowed to access and view full motion video clips that are a spatially filtered version of the video sequence. The filtered sequence retains enough information to make the material recognizable and attractive to the customer, but is not of sufficiently good image quality to tempt the customer to substitute the free browsing copy for the purchasable video source material. Browsing copy is extracted directly from Motion JPEG and MPEG video streams, in one embodiment of the invention, for providing substantial savings in the computational, storage and transmission costs. Temporal approximation and modelling of compressed video data is used to further enhance the performance of various embodiments of the present theft-resistant browser system invention.

In another embodiment of the invention, spatial filtering, in conjunction with temporal filtering, is used to provide visual information essential for browsing without tempting the customer to substitute the free browsing copy for the purchased source material. In yet another embodiment of the invention, payment means are provided for permitting a customer to pay a progressively higher fee, for progressively increased quality of the browsing copy.

Brief Description of the Drawings

Various embodiments of the present invention are shown and described herein with reference to the following drawings, in which like items are identified by the same reference designation, wherein: Fig. 1A is a pictorial presentation of a full video image having a resolution of 320 X 240 pixels;

Fig. IB shows the image of Fig. 1A lowpass filtered by (DC+2AC) a resolution of 80 X 60 pixels;

Fig. IC shows the image of Fig. 1A lowpass filtered via use of only DC coefficients at a resolution of 40 X 30;

Fig. 2A shows an Intra-coded image divided into 8 X 8 blocks;

Fig. 2B shows an enlargement of an 8 X 8 block of Fig. 2A;

Fig. 2C shows blocks of pixels within a video frame, including a reference block P_ref, motion vectors and original blocks P,, P₂, P₃, and P₄, for example;

Fig. 3A shows a video frame that has been lowpass filtered by (DC+2AC) ;

Fig. 3B shows the video frame of Fig. 3A after highpass filtering thereof; Fig. 4A shows a sequence of 20 frames from a newsprogram that have been lowpass filtered by (DC+2AC) ;

Fig. 4B shows the sequence of lowpass filtered frames of Fig. 4A after being highpass filtered; Fig. 5A shows an unfiltered video image of a frame from a newsprogram;

Fig. 5B shows a lowpass filtered frame of the original video image of Fig. 5A;

Fig. 5C shows a highpass filtered frame of the image of Fig. 5A;

Fig. 6 shows a block diagram of a system for one embodiment of the invention;

Fig. 7 shows a flowchart of the programming steps for a payment subsystem portion of the system of Fig. 6; Fig. 8 shows a simplified partial pictorial partial block diagram of the programming steps associated with a filtering subsystem of the system of Fig. 6;

Fig. 9 is a flowchart for the reconstruction of DC images for one embodiment of the invention; Fig. 10 is a flowchart for the reconstruction of (DC+2AC) images for one embodiment of the invention; and

Fig. 11 illustrates a DC+2AC image reconstruction of 2 X 2 block portions of a frame using a 2D inverse discrete cosine transform for each 2 X 2 block. Detailed Description of the Invention

In the present invention spatial filtering of each frame is provided as an alternative means of providing a subset of the original source material to the customer during browsing. Spatial filtering provides a mechanism similar to the browsing area of a magazine shop, which provides a space which allows the user to look through all the magazines to determine their content but is not sufficiently comfortable to encourage customers to read the magazines cover-to-cover without buying them. Spatial filtering provides a full-motion, full-length version of a video sequence which still protects the intellectual property rights of the copyright holders.

Research has been conducted by others to study the physiological limits on the amount of information required for video telephony and to characterize the information capacity and characteristics of the human visual system. In such prior research, tests were employed using as the base image sequences of films of American Sign Language (ASL) signing. Deaf subjects who understood ASL read the signing under various conditions of filtering, noise, etc.. By comparing what the subjects read from the sequences to the actual content, the experimenters accurately measured the intelligibility of the sequence to the entire visual understanding system.

These experiments showed the subjects could achieve very high recognition rates on sequences of very low resolution. Subjecting the video sequence to high-pass spatial filtering (i.e., edge highlighting) resulted in sequences which gave high recognition rates but which were judged by the subjects to be of low quality. The fact that the subjects' subjective evaluation of quality did not match their ability to extract information from the material suggests that filtering is a useful technique for video browsing. Making the customer slightly dissatisfied with the browsing copy is essential to inducing the customer to pay for the original source material. The present inventors recognized that one must balance the friendliness of the clip with .its attractiveness to theft, as described below. A filtering technique for one embodiment of the invention will now be described. The present browsing mechanism extracts the filtered video stream directly from Motion JPEG (Joint Photography Expert Group) or MPEG (Motion Picture Expert Group) video streams. Highpass and/or lowpass filtering may be used in various embodiments of the invention. Reduced lowpass-filtered images called DC images and DC+2AC images were tested by the present inventors, and found to be effective for theft-resistant browsing. The small size permits fast computation and fast data transfer, which in turn provides substantial savings in the computational, storage and transmission costs.

Motion JPEG and the I-frames in MPEG divide the original image into 8 X 8 blocks 5, for example, as shown in Fig. 2A, and apply a two-dimensional Discrete-Cosine Transform (DCT) to each block. The pixel values in a block are denoted by f(i,j), 0<i,j<7, and the DCT terms are denoted by C(i,j), 0<i,j<7. In MPEG-coded videos, motion compensation is applied to P-frames (predicted frames) and B-frames (bidirectional frames) ; the residue is coded using DCT. The DC term c(0,0) is related to the pixel values f(i,j) as follows:

which is 8 times the average intensity of the block, and wherein i and j are analogous to x and y, respectively.

A DC-image has one pixel per block in the original image, the pixel value being the average of the pixel values of the pixels in the corresponding block. A sequence of DC-images is called a DC-sequence. An original image 2 of size 320 X 240, its

(DC+2AC) image 3 of size 80 X 60, and DC image 4 of size 40 X 30, are shown respectively in Figs. 1A, IB and ic, for example. The (DC+2AC) image is comprised of blocks each of 2 X 2 pixels. Each block is obtained from the DC term C(0,0), and two AC terms C(1,0) and C(0,1) . The DC-image 4 may be thought of as a lowpass filtered, and a subsampled version of the original image 2. Although it is much smaller than the original image 2, the DC image 4 still retains a significant amount of information. However, details have been suppressed in DC image 4. In the example shown in Fig. IC of DC image 4, the words and numbers are no longer discernible, yet the overall information has been preserved. Embodiments of the invention for producing such images are given in detail below.

As is known to those of skill in the art, the DC image is the direct current equivalent to the average relative value of the 8 X 8 block. In the reconstruction of a block, as described below, each 8 X 8 block has sixty-four coefficients, of which one is the DC coefficient, and the remainder of the coefficients are called the AC coefficients. For example, Fig. 2B shows an 8 X 8 block portion 5 of the frame of Fig. 2A. As shown, the DC image portion or coefficient is in the upper left-hand corner, and AC coefficients AC01 and AC10 are adjacent to the DC coefficient, in this example. While forming the DC-image 4 from an uncompressed original image 2 requires 0(N²) operations per block (where N is the length of a block) , in the case of DCT-based compressed images, the DC term of a DCT block is a scaled version of the block's average value. When N=8, from equation (1), the average value is one-eighth the DC term of the DCT block. Thus, construction of a DC-image from each frame of Motion JPEG-code streams or I- frame of MPEG is easily accomplished.

For an intra-coded DCT frame (such as those in I-frames) , a DC-image is formed from the collection of scaled DC coefficients. The DC-image is reduced eight times in each dimension and may be thought of as a lowpass filtered and subsampled version of the original image. Although it is much smaller than the original image, it still retains significant amount of information. However, details have been suppressed. To construct a DC-image from P-frames and B-frames directly from compressed streams, approximation algorithms have been developed to construct images that are very close to the actual DC images. The construction is described below.

DC images can be constructed directly from a compressed MPEG stream. The following example focuses upon P-frames and B- frames. With reference to Fig. 2, let P_ref be the current block of interest, and P₁₇ P₂, P₃ and P₄ be the four original neighboring blocks from which P_ref is derived. Also, let the 2-D DCT of a block P be DCT(P) , and its (i,j) component be (P)_s. The DC value of DCT(P_ref) as derived by Yeo and Liu in their above-mentioned paper follows below. The DCT coefficients of P_ref are as follows:

DCT{P_IBt) = ∑ DCT(S_i:L) DCT{P_i) DCT{S_i2) (2 ) i-i where S-,- are matrices of the form

As shown below in Table 1, there are four possible locations of the subblock of interest: lower right, lower left, upper right, and upper left, giving different values for S*_j as follows, where h* and w* are the height and width of the subblock of interest in P*.

Table 1

The actual DC value of DCT (P_ref) is then as follows:

(DCTiP_Iθf) ) ι (DCTlPi) ) * (4)

where <*mi = (DCT(S_a) )_to X (DCT(S_i2))_] The matrices show that the weight ^ωoo of is (h; X w*)/64, and that (OCT (Ε>_nf) ) _<„ can be accurately approximated as shown in equation (5) below:

(Dcτ(p_zβ£) ) hwωcπp,)) ₀₀ <⁵⁾

The reconstruction of (DCT(P_ref))oo by equation (5) is a first-order approximation. Such approximation can be applied to both P-frames and B-frames, but because P-frames can potentially be used to code other P-frames and B-frames, error propagation will occur. However, the error will be small. The net effect is increasing blurring of the reconstructed DC image as the distance of the image from the nearest previous I-frame increases. Even with the blurring, the global features are well preserved.

Further approximating equation (5) by the following equation (6) is possible:

(DCT(P_zef) ) ₀₀= (DCTiP ) ₀₀, where w_i x h_i ≥ 16 (6)

In this example, the DC value of the block with largest subblock of interest is used. This provides a zero-order approximation. The DC sequence constructed by this approximation is still viable for the purposes of the present invention.

The DC+2AC reduced image is reduced four times in each dimension. For an intra-coded block, it is formed by taking an 2 X 2 inverse DCT using only the DC and two AC coefficients, c(0,l) and c(l,0). To reconstruct such images from P and B frames, approaches similar to first-order approximation used for DC images are used. Only the DC and two AC coefficients and two AC coefficients are used in the anchor frames to reconstruct the DC and two AC coefficients of the present frames. Higher order terms in equation (2) are ignored. In this case, reconstructions are as follows in equations (7) , (8) , and (9) :

(7)

(DCT(P_τef) ) ₀₀ - ∑ DCT(S ) _0mDCT(S_i2) u CTiP ) _ml i=ι l +m≤l

(8)

(DCT(P_∑ef) ) _Q1 = ∑ DCTiS _0πpCT(S_i2) _u (DCTiP,) ) _ml

2-1 \l +m≤l

ωcτt.p„_f) ) „ •

Each coefficient requires twelve multiplications and twelve additions. The 2 X 2 inverse DCT operations involve only additions and subtractions and no multiplications.

Formation of the DC image is a first step in spatial filtering of the image sequence for browsing. One can perform edge detection on the DC image to further filter the images, preserving information important for recognition while eliminating information which makes the images more pleasant to view. The computational complexity is extremely low because only a small portion of data is involved. In another embodiment of the invention, a Sobel operator is used as the high-pass filter on the DC sequences. While other more sophisticated schemes are possible, the Sobel operator is used for its simplicity.

The Sobel edge detector is defined by two masks ≤_! and S₂:

Masks S_j and S₂ are used to detect horizontal and vertical edges, respectively. Denoting the response to S* by g*, the total response is formed as [--—— !—IL] . Implementation of such operations involve only shifts and additions. In Fig. 3A, a typical (DC+2AC) frame 6 is shown. In Fig. 3B, the frame 8 illustrates the frame 6 of Fig. 3A after highpass filtering. From the results, one can see that the essence of the video is still preserved for understanding of the image information conveyed, while permitting effective and yet theft- resistant browsing of the materials via the filtered frames 10 of Fig. 3C. Similar comments apply to Figs. 4A and 4B showing lowpass and highpass filtered versions, 10, 12, respectively, of twenty frames of a news program; and to Figs. 5A, 5B, 5C showing original, lowpass and highpass versions 14, 16, 18, respectively, of a newscaster frame.

As discussed in previous sections, the first-order and zero- order approximations applied in the reconstruction of (DCT(P_ref) )₀₀ will further blur the reconstructed DC image as the distance of the image from the nearest previous I-frame increases. Experimental results have shown that reduced video streams processed by these approximation techniques not only can offer additional protection means, but also are able to maintain their effectiveness for browsing purposes, while at the same time its computational, storage and transmission requirements are minimal. In Fig. 6, a browsing system configuration for another embodiment of the invention is shown. The browsing system 20 includes a video monitor 22, a browsing control system 24, a payment system 26, a filtering system 28, and a computer memory device 30. The browsing control 24 can be provided by the keyboard of a computer having incorporated therein a mouse or joystick type control, for example. The payment system 26 and filtering system 28 can be provided through appropriate programming of an associated computer (not shown for the sake of simplicity) , for example. The memory device 30 can be provided, for example, by a disk or other storage device, for storing a collection of video programs in compressed motion JPEG or MPEG formats, for example. The software based payment system 26, and filtering system 28, are described below in greater detail relative to the flowcharts of Figs. 7 and 8, respectively.

In Fig. 7, the payment system 26, in this example, is a software based system driven by a user operating the browsing control or keyboard 24 in typically a menu driven format. The user, when first signing onto the system, provides the user's account number as shown in step 28, whereafter the next step is for the system to check the account number against valid account numbers, to determine whether the users account number is valid, as shown in decision step 30. If the account is not valid, the system outputs a signal which displays to the user on monitor 22 the invalid account status, and denies the user access to the video programs. If the account is confirmed to be valid in step 30, the user is then requested in step 32, via prompts on monitor 22, to select the particular available video program of interest, and also to select the level of quality of the browsing copy to be reviewed of the selected video program. The higher the selected quality for the browsing copy, the higher the payment or fee that will be charged for granting the user access to the selected video program at the selected quality level. After these selections are made in step 32, payment information is gathered in an account update step 34, and the user is provided with either or both of a display of the users account on monitor 22, and/or a printout via a printer (not shown for sake of simplicity) . Also, from selection step 32, an output signal 36 is provided for the index of the selected video program, and another output signal 38 is provided for indicating the index of the selected level of quality for the browsing copy of the selected program 36.

With reference to Fig. 8, the index of selected program 36 is used by the computer for retrieving the associated video program from the memory 31 storing the various video databases or available programs, and providing it to filtering system 28, as shown. The software based spatial filtering system 40 is responsive to the index of selected quality 38 for adjusting the degree of filtering to the index 38. As shown, depending upon the level of the index of selected quality 38, the spatial filtering system 40 may provide the maximum amount of filtering through use of only the DC coefficient and/or highpass filtering 44, or perhaps a lesser degree of filtering through use of coefficients including (DC+2AC) , all the way to the extreme of no filtering for providing the original resolution. Regardless, in step 42, the selected video program is reconstructed through use of the spatial filtering level selected, and the resultant browsing video copy is supplied to the user via display of the browsing material on monitor 22.

Note that in step 44, a decision is made whether highpass filtering of the lowpass filtered video program is to be made. If not, the lowpass filtered video program is passed through to the user for display on monitor 22. If highpass filtering is required, the lowpass filtered video is passed through highpass filter 46 (typically edge detection via a Sobel filter) , and then to the user or monitor 22.

In Fig. 9, a flowchart for reconstruction step 42 for DC images, includes steps 48, 50, 52, and 54. In step 48 each new frame is lowpass filtered via the filter selection of step 40 being DC, in this example. Next, in step 50 a determination is made as to whether the frame is intra coded. If the answer is "yes", step 54 is entered for extracting DC coefficients from the current frame, and proceeding to step 44. If the answer is "no", step 52 is entered for extracting DC coefficients from anchor frames, and then proceeding to step 44.

An alternative reconstruction step 42 is shown in Fig. 10 for reconstructing [DC+2AC] images. In step 56, each new frame is lowpass filtered via a filter selection of [DC+2AC] , in this example. Next, in step 58 a determination is made as to whether the current frame is intra coded. If not, step 60 is entered for extracting DC, AC01, and AC10 coefficients from anchor frames, and then proceeding to step 64 for making a 2D inverse discrete transform for each 2 X 2 block, in this example. The resultant [DC+2AC] images are then passed to step 44. If the current frames is found to be intra coded in step 58, step 62 is entered for extracting DC, AC01, and AC10 coefficients from the current frame, and followed by step 64 for proceeding as previously indicated. In Fig. 11, the DC+2AC image reconstruction for a 2 X 2 block in an upper left corner is illustrated.

Although various embodiments of the invention are shown and described herein, they are not meant to be limiting. Various modifications to these embodiments may occur to those of skill in the art. For example, the embodiments of the invention described above generally use DC-sequences extracted from MPEG or Motion-JPEG based compressed video. However, the proposed methods of the invention can also be applied to other popular compression techniques such as subband or wavelet coding. In addition, temporal subsampling of the video sequence can be an alternative to further reducing the data to be transmitted.

Claims

What is Claimed is:

1. A method for providing theft resistant video browsing of video material on a video monitor comprising the steps of:

storing said video material in a memory device;

permitting a user to select a particular video program from said stored video material for presentation on said video monitor;

filtering the user selected said video program before presentation to reduce the quality of the video content thereof to a level sufficient for the user to browse the video program to determine whether to purchase, rent, or license said video program, while insuring the level of quality of the video program is insufficient to permit the user to steal a usable copy thereof, or satisfy a need to see the video program; and

presenting the filtered user selected video program to said user by playing it on said video monitor.

2. The method of claim 1, further including the step of compressing said video material into a compressed format before the step of storing.

3. The method of claim 1, further including before said storing step, the step of converting said video material into a compressed motion format, thereby permitting the user to browse a full motion video program.

4. The method of claim 1, further including before said storing step, the step of compressing said video material into a compressed motion JPEG format.

5. The method of claim 1, further including before said storing step, the step of compressing said video material into a compressed motion MPEG format.

6. The method of claim 1, further including before said user selection permitting step the steps of:

requesting a user to provide the user's account number; and

validating the user's account number, whereby if the account number is validated, the user is requested to carry out said user selection permitting step, but if the account number is found to be invalid, the user is denied access to said user selection permitting step.

7. The method of claim 6, whereby after a user's account is validated in said validating step, said user selection permitting step further includes the steps of:

providing an index of selected program signal indications of the user's video program selection;

requesting the user to select in accordance with a pricing schedule a level of quality for the video program selected for browsing; and

providing an index of selected quality based upon the user's quality level selection.

8. The method of claim 7, wherein said step of filtering further includes the step of adjusting the magnitude of spatial filtering of said video program to correspond to the index of selected quality, whereby the degree of filtering ranges from maximum filtering for the lowest index of quality, to no filtering or original resolution for the selected video program for the highest index of quality.

9. The method of claim 1, wherein said step of filtering includes lowpass spatial filtering.

10. The method of claim 1, wherein said step of filtering includes highpass spatial filtering.

11. The method of claim 1, wherein said step of filtering includes the steps of:

lowpass spatial filtering said video program; and highpass spatial filtering the lowpass filtered video program.

12. The method of claim 8, wherein said step of filtering includes lowpass filtering.

13. The method of claim 12, wherein the range of lowpass filtering consists of DC, (DC+2AC), ...(original resolution).

14. The method of claim 10, wherein said step of highpass spatial filtering consists of using a Sobel operator as the highpass filter.

15. The method of claim 11, wherein said step of highpass spatial filtering consists of using a Sobel operator as the highpass filter.

16. The method of claim 1, further including before said storing step, the step of temporal subsampling of said video program for obtaining said video material to be stored in said storing step.

17. The method of claim 3, wherein said converting step includes subband encoding of said video material.

18. The method of claim 3, wherein said converting step includes wavelet encoding of said video material.

19. The method of claim 3, further including the step of temporal subsampling.

20. The method of claim 17, further including the step of temporal subsampling of said subband encoded video material.

21. The method of claim 18, further including the step of temporal subsampling of said wavelet encoded video material.

22. The method of claim 8, wherein if the index of selected quality for spatial filtering dictates use of DC coefficients or images, reconstruction of a filtered image or frame includes the steps of:

determining whether a current frame is an intra coded frame;

extracting DC coefficients from anchor frames if the frames are not intra coded; and

extracting DC coefficients from a current frame if the current frame is intra coded, to obtain a browsing video program.

23. The method of claim 8, wherein if the index of selected quality for spatial filtering dictates use of [DC+2AC] coefficients or images, reconstruction of a filtered image or frame includes the steps of:

determining whether a current frame is an intra coded frame;

extracting DC, AC01, and AC10 coefficients from anchor frames if the frames are not intra coded;

extracting DC, AC01, and AC10 coefficients from a current frame if the current frame is intra coded; and

transforming each extracted DC, AC01, and AC10 coefficient via a 2D inverse discrete cosine function for each 2 X 2 block of the associated frame or image, to obtain a browsing video program.