B. Girod. The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals. pages 178--187. Proc. of SPIE Vol. 1077, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....are significantly reduced, but the PSNR drops as well. tem has a spatial masking effect which can cause decreased sensitivity to errors very close to edges. Indeed, the visual masking effect has been exploited in previous compression techniques, such as the video coding algorithm described in [17]. We desire, though, to exploit geometric modeling to the extent that it actually improves PSNR, not just perceptual quality. 3.2.3 Additional approaches While the above approach focused on exploiting geometrical information available to the encoder, it is also possible to exploit information ....

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals," in Proc. SPIE Human Vision, Visual Processing, and Digital Display, 1989, vol. 1077, pp. 178--187.

.... the HVS frequency sensitivity and masking are closely related, in that both effects suggest we are less sensi tive to noise in high frequency, which means that if we take advantage of the frequency sensitivity, this will directly reduce the amount of redundancy that masking can take advantage of [26, 37]. Also, the masking effect is very com plex and cannot always be reliably modeled. It is more pronounced on noisy or random backgrounds than it is on regular, easily learned backgrounds [88] Baseband Figure 1.6: Typical cortex transform decomposition in the frequency domain Part of the ....

B. Girod. The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals. In SPIE, Human Vision, Visual Processing, and Display, volume 1077, pages 178-187, 1989.

....models for signal compression is presented in [22] A different image compression model has been developed in [2,23] This model uses both frequency sensitivity, and spatial masking based on an image s edges. The spatial masking is a modified version of the spatial masking model presented in [24]. The main principle of the spatial masking is that edges in an image are able to mask signals of much greater amplitude than regions of near constant intensity. For a given image, a tolerable error level value (TEL) may be formed for each pixel. This quantity is similar in concept to the JND ....

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals," Proceedings of the SPIE International Conference on Human Vision, Visual Processing, and Digital Display, vol. 1077, 1989, pp. 178-187.

....visual models for signal compression is presented in [6] A different image compression model has been developed in [11] This model uses both frequency sensitivity, and spatial masking based on an image s edges. The spatial masking is a modified version of the spatial masking model presented in [12]. The main principle of the spatial masking is that edges in an image are able to mask signals of much greater amplitude than regions of near constant intensity. For a given image, a tolerable error level (TEL) may be formed for each pixel. This quantity is similar in concept to the JND. The TEL ....

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals," Proceedings of the SPIE International Conference on Human Vision, Visual Processing, and Digital Display, vol. 1077, 1989, pp. 178-187.

....the patterns. This is equivalent to knowing the secret key. We try to answer the question of how many hash bits can be changed using the knowledge of projections by making imperceptible changes to the pixel gray levels. The maximal allowable changes were determined by the masking model of Girod [15]. The constraints imposed by the masking model also constrain the maximal possible changes in the projections c = B.P . Consequently, not all hash bits can be flipped. The maximal allowable change for the projection ck is determined by the expression EuIPu(k)I j, where y is the masking value ....

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178 187, 1989.

....adding. Zhu et al. 4] propose two techniques based on spatial and frequency masking. Their watermark is guaranteed to be perceptually invisible, yet it can detect errors up to one half of the maximal allowable change in each pixel or frequency bin depending on whether frequency [5] or spatial [6] masking is used. The image is divided into blocks and in each block a secret random signature (a pseudo random sequence uniformly distributed in [0,1] is multiplied by the masking values of that block. The resulting signal depends on the image block and is added to the original block quantized ....

....60 55 14 13 16 24 40 57 69 56 14 17 22 29 51 87 80 62 18 22 37 56 68 109 103 77 24 35 55 64 81 104 113 92 49 64 78 87 103 121 120 101 72 92 95 98 112 100 103 99] The quantized values are further binary encoded. The bit lengths of their codes (including the signs) are shown in matrix L L=[7 7 7 5 4 3 2 1 7 6 5 5 4 2 1 0 6 5 5 4 3 1 0 0 5 5 4 3 1 0 0 0 4 4 3 1 0 0 0 0 3 2 1 0 0 0 0 0 2 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0]. Coding based on L will guarantee that the first 11 coefficients from each block will be coded using exactly 64 bits. In the rare event when the i th DCT coefficient has absolute value is larger than 2 Li , only this maximum available value will be encoded. Step 3 (Encrypting and embedding) ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178-187, 1989.

....are used in highquality, low bit rate coding algorithms to further reduce bit rates [17] The masking models reviewed here are based on image models. A detailed discussion of the models may be found in [18] We are currently developing a watermarking algorithm that takes temporal masking (e.g. [19]) into account. A. Frequency Masking Our frequency masking model is based on the knowledge that a masking grating raises the visual threshold for signal gratings around the masking frequency [20] The model we use [18] based on the discrete cosine transform (DCT) expresses the contrast ....

....rule of the form (2) to sum up the masking effects from all the masking signals near If the contrast error at is less than the model predicts that the error is invisible to human eyes. B. Spatial Masking Our spatial masking model is based on the threshold vision model proposed by Girod [19]. The model accurately predicts the masking effects near edges and in uniform background. Assuming that the modifications to the image are small, the upper channel of Girod s model can be linearized [18] to SWANSON et al. MULTIRESOLUTION SCENE BASED VIDEO WATERMARKING USING PERCEPTUAL MODELS 543 ....

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals," in Proc. SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, 1989, pp. 178--187.

....masking ) but also before. This backward masking may be explained as the result of the variation in the latency of the neural signals in the visual system as a function of their intensity [4] The opposite of temporal masking, temporal facilitation, can occur at low contrast discontinuities [38]. So far, the above mentioned temporal effects have received much less attention in the video coding community than their spatial counterparts. In principle, temporal masking can be taken into account with a contrast gain control model as in Equation (3) by adding a time dependency to the pooling ....

....effects have received much less attention in the video coding community than their spatial counterparts. In principle, temporal masking can be taken into account with a contrast gain control model as in Equation (3) by adding a time dependency to the pooling function h, as demonstrated by Girod [38]. Watson [118] recently outlined a PVQA system 18 that models forward masking effects by means of a masking sequence produced by passing the reference through a low pass filter. 10 Pattern Adaptation Pattern adaptation adjusts the contrast sensitivity of the visual system in response to the ....

B. Girod: "The information theoretical significance of spatial and temporal masking in video signals." in Proc. SPIE , vol. 1077, 178--187, Los Angeles, CA, 1989.

....of f . If the contrast error at frequency f is less than c(f ) the model predicts that the error is perceptually invisible. Similar masking effect exists around an edge in spatial domain. The model of spatial masking is a modified model from the threshold vision model proposed by Girod [6]. In the model, the processing channel is linearized under the assumption that the perceptual error at threshold vision is small. The perceptual threshold at each pixel is found reversely from the last stage to the first stage. For details, readers are referred to [3] 3. EMBEDDING SIGNATURE ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals," SPIE Vol. 1077 Human Vi- sion, Visual Processing, and Digital Display, pp. 178-187, 1989.

....adding. Zhu et al. 4] propose two techniques based on spatial and frequency masking. Their watermark is guaranteed to be perceptually invisible, yet it can detect errors up to one half of the maximal allowable change in each pixel or frequency bin depending on whether frequency [5] or spatial [6] masking is used. The image is divided into blocks and in each block a secret random signature (a pseudo random sequence uniformly distributed in [0,1] is multiplied by the masking values of that block. The resulting signal depends on the image block and is added to the original block quantized ....

....19 26 58 60 55 14 13 16 24 40 57 69 56 14 17 22 29 51 87 80 62 18 22 37 56 68 109 103 77 24 35 55 64 81 104 113 92 49 64 78 87 103 121 120 101 72 92 95 98 112 100 103 99] The quantized values are further binary encoded. The bit lengths of their codes (including the signs) are shown in matrix L L=[7 7 7 5 4 3 2 1 7 6 5 5 4 2 1 0 6 5 5 4 3 1 0 0 5 5 4 3 1 0 0 0 4 4 3 1 0 0 0 0 3 2 1 0 0 0 0 0 2 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0]. Coding based on L will guarantee that the first 11 coefficients from each block will be coded using exactly 64 bits. In the rare event when the i th DCT coefficient has absolute value is larger than 1 2 i L , only this maximum available value will be encoded. Step 3 (Encrypting and ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178--187, 1989.

....coefficients are modified by a larger amount. Cox et al. suggest the value a = 0.1 obtained empirically. In [Fri2] the author studied the visibility of the watermark using a model of the human visual system. The model used was a simplified version of a linearized spatial masking model of Girod [Gir1] This model accurately describes the visibility of small changes in uniform areas and around edges. It was found that the value of a = 0.1 introduces artifacts that are fairly visible even to an inexperienced observer. Over 15 of pixels exhibited visible changes after watermark insertion. A ....

....using an inverse DCT. Since the frequency masking model is based on idealized assumptions of two sinusoidal gratings on a uniform background, its accuracy for real images may not be sufficient. To guarantee perceptual invisibility of the changes, the linearized spatial masking model of Girod [Gir1] is used to provide feedback whether or not the changes are visible. If they are, the masking values M ij are multiplied by a factor less than one and the process is repeated till no visible changes are produced. The security of the scheme is in the secret autohor s ID that is used to produce the ....

[Article contains additional citation context not shown here]

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178-- 187, 1989.

....30 of DCT coefficients by adding the spread spectrum signal. The amplitude of the added signal can be adjusted to achieve balance between watermark visibility and robustness. We set the amplitude equal to 13 (we used the symmetric form of DCT) Using the linearized spatial masking model of Girod [6] without the temporal aspect, the watermark was visible for 0.17 of all pixels. The detection of the watermark proceeds by blocks. For each block, M bits are extracted and the block is DCT transformed. Then, the spread spectrum signal is synthesized using the camera ID and the PRNG. Total M ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178--187, 1989.

....knm . In other words, instead of using the same limits k min ; k max for all image pixels, one can use the corresponding value of the JND profile for this purpose . Gammaj nd(m; n) kmn jnd(m; n) 36) We have experimented with two different JND evaluation methods that have been proposed in [28], 29] Unfortunately, both JND evaluation methods did not give satisfactory results (i.e. the predicted invisible distortions were rather noticeable) The cause of this failure lies in the fact that both these models depend on a number of parameters (image observer distance, background ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", SPIE vol 1077, Human Vision, Visual Processing, and Digital Display, pp. 178-187, 1989.

....on Perceptual Watermarking Zhu et al. 5] propose two techniques based on spatial and frequency masking. Their watermark is guaranteed to be perceptually invisible, yet it can detect errors up to one half of the maximal allowable change in each pixel or frequency bin depending on whether spatial [6] or frequency [7] masking is used. The image is divided into blocks and in each block a secret random signature (a pseudo random sequence uniformly distributed in [0,1] is multiplied by the masking values of that block. The resulting signal depends on the image block and is added to the original ....

....of DCT coefficients by adding the spread spectrum signal. The amplitude of the added signal can be adjusted to achieve a balance between watermark visibility and robustness. We set the amplitude equal to 13 (we used the symmetric form of DCT) Using the linearized spatial masking model of Girod [6] without the temporal aspect, the watermark was visible for 0.17 of all pixels. 2.2 Watermark Detection The detection of the watermark proceeds by blocks. For each block, M bits are extracted and the block is DCT transformed. Then, the spread spectrum signal is synthesized using the camera ID ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178--187, 1989.

.... is used) To exploit the background luminance masking effect, the quantization step size for each block is adapted to the local impairment visibility threshold by linearly scaling q max [6] as shown in Figure 1(b) This scaling is a linear approximation of the impairment visibility curve in [7]. The background luminance level is estimated by the mean intensity value of the corresponding predicted block in the previous frame. Using predicted blocks avoids transmission of the quantization step size for each block; instead, only q max needs to be transmitted. There are two advantages over ....

Bernd Girod, "The information theoretical Significance of spatial and temporal masking in video signals, " in SPIE Visual Commun. Image Processing, vol. 1077, pp. 178-187, 1989.

....error at f is less than c(f ) the model predicts that the error is invisible to human eyes. Spatial masking refers to the situation that an edge raises the perceptual threshold around it. The model used here is similar to our image coding model [8] which is based on a model proposed by Girod [9]. In our approach, the upper channel of Girod s model is linearized under the assumption of small perceptual errors. The model gives the tolerable error level for each pixel in the image. 4. SPATIAL DATA HIDING In Fig. 1, we show our first data hiding technique which uses spatial masking to ....

B. Girod, "The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals," in Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178--187, 1989.

No context found.

B. Girod. The Information Theoretical Significance of Spatial and Temporal Masking in Video Signals. pages 178--187. Proc. of SPIE Vol. 1077, 1989.

No context found.

B. Girod, "The information theoretical significance of spatial and temporal masking in video signals", Proc. of the SPIE Human Vision, Visual Processing, and Digital Display, vol. 1077, pp. 178--187, 1989.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC