Global motion compensation concerns the estimation and compensation of motion characteristics that affect the whole frame, as for example in video clips shot using a hand-held camera. In the example shown in the Figure 1 below, matched features in the two images below indicate a translation and rotation motion between the two pictures. In general, the key steps involved in estimating global motion comprise identifying features in both images, matching the identified features, and estimating global motion parameters based on the matched features.
The general motion model is given by:
where
-
Affine projection:
$h_{31} = h_{32} = 0, h_{33}=1$ . This transformation preserves parallelism and has six parameters to estimate. -
Rotation-zoom projection:
$h_{31} = h_{32} = 0, h_{33}=1; h_{11}=h_{22};h_{12}=-h_{21}$ , which corresponds rotation + scaling. This transformation preserves angles and has four parameters to estimate. -
Translation:
$h_{31} = h_{32} = 0, h_{33}=1; h_{11}=h_{22}=1;h_{12}=h_{21}=0$ . This transformation preserves orientation and size and has two parameters to estimate.
The global motion estimation involves two main steps. The first step concerns feature matching where the objective is to identify features that are present in both the source and reference pictures. The second step concerns model identification, where the identified features are used to estimate the motion model parameters. In SVT-AV1, the global motion parameters are computed for each reference frame using feature matching followed by applying the random sample consensus (RANSAC) algorithm. The estimated parameters are sent in the bitstream.
To identify features that are common to both the source and reference pictures, the features from the Accelerated Segment Test (FAST) algorithm are used as a feature detector. The Fast algorithm identifies corner points by examining a circle of 16 pixels (Brensenhan circle of radius 3) around the pixel p of interest. If out of the 16 pixels, 12 contiguous pixels all have values above the pixel p by at least a given threshold or all have values below that of p by at least a given threshold, then the pixel is considered a feature (corner point) in the image. Such features are robust to motion and brightness changes. Once features on the source frame and on the reference frame are identified, feature matching is performed by computing the normalized cross-correlation function between the two sets of features. A feature (i.e. corner point) is selected if:
-
The feature on the reference frame is located within a pre-specified distance from the feature in the source frame.
-
The correlation between the point in the reference frame and that in the source frame is highest.
The model is identified based on the matched feature points from the feature matching step. A least squares estimation is performed to compute the model parameters using the matched feature points. The RANSAC (Random Sample Consensus) algorithm is used in the estimation. The algorithm minimizes the impact of noise and outliers in the data. The set of parameters to be estimated depends on the motion model (Translation, rotation-zoom, affine) specified. The identified parameters are included in the bitstream.
The RANSAC algorithm finds model parameters that yield the best match to the motion of the identified features. The steps involved in the algorithm are as follows:
-
A small number of matched features (corner points) are used in the model parameter estimation (as dictated by the number of parameters to estimate).
-
The remaining features are used to evaluate the fitness of the model by counting the number of those matched features where the model yields a small error (inliers). The remaining tested features are considered outliers.
-
Steps 1 and 2 are repeated based on another small set of matched features and the number of resulting outliers is recorded.
-
The process stops when the number of outliers is below a specified threshold.
Input to Motion Estimation: Input frames of the stream. Outputs of Motion Estimation: Estimated global motion models per frame with their references. Input to Mode Decision: Estimated global motion models. Outputs of Mode Decision: Encoded frame with global motion encoded blocks if they provide a cost advantage.
Table 1 below summarises the invoked functions when global motion is enabled. The process where each function is invoked is also indicated as well as a brief description of each function.
| Process | Function | Purpose |
|---|---|---|
| Picture Decision Process | set_gm_controls | Set global motion controls |
| Motion Estimation Process | perform_gm_detection | Detect whether a global motion may be identified based on the uniformity of the motion vectors produced by the normal motion estimation search |
| Motion Estimation Process | global_motion_estimation | Perform global motion estimation search |
| Mode Decision Configuration process | set_global_motion_field | Map the global motion information generated in EbMotionEstimationProcess to EbEncDecProcess |
| Mode Decision Process | inject_global_candidates | Inject global motion as a mode candidate to the mode decision |
The global motion data flow is summarized in the Figure 2 below.
The main algorithmic components of the global motion feature are the estimation component which takes place in the motion estimation process, and the injection and processing component which takes place in the Mode Decision process(injection and processing).
This process is executed by the global_motion_estimation function. This function
is called only for the first segment of each frame in the motion_estimation_kernel
but it computes the global motion for the whole frame. The function involves a loop
that runs over all reference frames.
To compute the global motion between two frames, the FAST features of the reference
frames are extracted and matched to those of the current frame in the svt_av1_fast_corner_detect
function, thanks to the fastfeat third-party library. The svt_av1_fast_corner_detect function
is first called to determine the features in the source picture. Then it is called again
from the function svt_av1_compute_global_motion to determine the features in the reference picture.
Once the features have been extracted, they are matched. This is done in the
svt_av1_determine_correspondence function by two nested loops over the features of the
reference frame and the current frame. A current frame feature is matched to a reference
frame feature that maximizes their cross-correlation computed by svt_av1_compute_cross_correlation_c.
However, the match is kept only if the cross-correlation is superior to the THRESHOLD_NCC
threshold multiplied by the variance of the current feature patch.
The matched feature positions are further refined in the improve_correspondence function.
This function performs a double iteration to look for the best match in a patch of size
SEARCH_SZ located around the previously found match position.
The rotation-zoom and affine global motion models are tested with the RANSAC algorithm
by the ransac function. This function takes as argument three function pointers:
is_degenerate, transformation and projectpoints. They are set according to the type
of transformation that is estimated.
The minimum number of transformation estimation trials is defined by the MIN_TRIALS macro.
For each trial, the algorithm selects random feature match indices with the get_rand_indices function.
It first checks if the current match selection does not lead to a degenerated version of
the transformation with the is_degenerate function pointer. The parameters of the
transformation are then estimated by the find_transformation function pointer.
The positions of the feature matches that have not been used to compute the transformation
parameters are projected with the projectpoints function pointer. Finally, the number of
inliers and outliers of the current transformation are counted. A feature match is considered
as an outlier if its distance with its position calculated with the transformation is
superior to the INLIER_THRESHOLD macro.
The parameters of the top RANSAC_NUM_MOTIONS transformations that have the greatest
numbers of inliers and smallest position variance are kept. These transformations are
then ranked by their number of inliers and their parameters are recomputed by using
only with the inliers.
The transformation parameters are refined in the svt_av1_refine_integerized_param function.
It uses the svt_av1_warp_error function to estimate the error between the reference frame
and the current frame in order to select the model with the smallest error.
As saving global motion parameters takes space in the bit stream, the global motion model
is kept only if the potential rate-distortion gain is significant. This decision is made
by the svt_av1_is_enough_erroradvantage function thanks to the computed frame error, the storage
cost of the global motion parameters and empirical thresholds.
The AV1 specifications define four global motion types:
-
IDENTITY for an identity model,
-
TRANSLATION for a translation model,
-
ROTZOOM for a rotation and zoom model,
-
AFFINE for an affine model.
Each block that is 8x8 or larger in size can be a candidate for local or
global warped motion. For each block, we insert in the
inject_inter_candidates function global motion candidates for the
simple and compound modes for the LAST_FRAME and the BWDREF_FRAME
frame types. The compound mode implementation only mixes global warped
motions for both references.
To identify global warped motion candidates, the
warped_motion_prediction function has been modified to support the
compound mode for warped motions for the case where high bit-depth is
enabled and for the case where it is not.
The two main steps involved in MD are the injection of GLOBAL and GLOBAL_GLOBAL candidates, and the processing of those candidates through MD stages. The conditions for the injection of GLOBAL candidates are as follows: For the case where downsample_level <= GM_DOWN:
- The global motion vector points inside the current tile AND
- (((Transformation Type > TRANSLATION AND block width >= 8 AND block height >= 8) OR Transformation type <= TRANSLATION))
Otherwise, only condition 1 above applies.
The conditions for the injection of GLOBAL_GLOBAL candidates are as follows:
For the case where downsample_level <= GM_DOWN:
- Is_compound_enabled (i.e. compound reference mode) AND
- allow_bipred (i.e. block height > 4 or block width > 4) AND
- (List_0 Transformation type > TRANSLATION AND List_1 Transformation type > TRANSLATION))
Otherwise, only conditions 1 and 2 above apply.
It should be noted that for the case of compound mode prediction, only GLOBAL_GLOBAL candidates corresponding to compound prediction modes MD_COMP_AVG and MD_COMP_DIST are injected.
The three main functions associated with the injection of GLOBAL_GLOBAL candidates are
precompute_intra_pred_for_inter_intra, inter_intra_search and determine_compound_mode.
The first two are related to the generation of inter-intra compound candidates. The third
is related to the injection of inter-inter compound candidates.
With respect to ranking the global motion candidates, the current implementation
uses the specific class (CAND_CLASS_8) that adds a dedicated path for those candidates.
This allows some of the those candidates to survive until the last and most costly stage
of the mode decision process.
Different quality-complexity tradeoffs of the global motion algorithm can be achieved by manipulating a set of control parameters that are set in the gm_controls() function. These control parameters are set according to the flag gm_level which is set in the picture decision process according to the encoder preset. The different parameters that are controlled by the flag gm_level are described in Table 2 below.
| Flag | Level (Sequence/Picture) | Description |
|---|---|---|
| enabled | Picture | Enable/Disable global motion knob |
| identiy_exit | Picture | 0: Generate GM params for both list_0 and list_1, 1: Do not generate GM params for list_1 if list_0/ref_idx_0 is identity. |
| rotzoom_model_only | Picture | 0: Use both rotzoom and affine models, 1: Use rotzoom model only |
| bipred_only | Picture | 0: Inject both unipred and bipred GM candidates, 1: Inject only bipred GM candidates. |
| bypass_based_on_me | Picture | Bypass global motion search based on the uniformity of motion estimation MVs. 0/1: Do not bypass/Bypass GM search. |
| use_stationary_block | Picture | 0: Do not consider stationary_block info at me-based bypass, 1: Consider stationary_block info at me-based bypass (only if bypass_based_on_me=1) |
| use_distance_based_active_th | Picture | Active_th is the threshold used to decide on the uniformity of MVs from motion estimation. 0: Use default active_th, 1: Increase active_th based on distance to ref (only if bypass_based_on_me=1) |
| params_refinement_steps | Picture | Specify the number of refinement steps to use in the GM parameters refinement. |
| downsample_level | Picture | GM_FULL: Exhaustive search mode. GM_DOWN: GM search based on down-sampled resolution with a down-sampling factor of 2 in each dimension. GM_TRAN_ONLY: Translation only using ME MV |
The generated global motion information may be used in all or some of the mode decision Partitioning Decision (PD) passes. The injection of global motion candidates in MD is controlled by the flag global_mv_injection.
The global motion parameters are written in the bitstream for each encoded frame with their corresponding references.
Boolean parameters encode the type of global motion models among the four available: IDENTITY, TRANSLATION, ROTZOOM or AFFINE (See Table 3).
| Frame level | Values | Number of bits |
|---|---|---|
| is_global | {0, 1} | 1 |
| is_rot_zoom | {0, 1} | 1 |
| is_translation | {0, 1} | 1 |
Depending on the model complexity, several parameters are also encoded (See Table 4). Each of those parameters corresponds to an entry in the affine transformation matrix.
| Frame level | Number of bits |
|---|---|
| Global motion parameters: | Up to 12 |
| 0 parameter for IDENTITY | Up to 12 |
| 2 parameters for TRANSLATION | Up to 12 |
| 4 parameters for ROTZOOM | Up to 12 |
| 6 parameters for AFFINE | Up to 12 |
The feature settings that are described in this document were compiled at v1.3.0 of the code and may not reflect the current status of the code. The description in this document represents an example showing how features would interact with the SVT architecture. For the most up-to-date settings, it's recommended to review the section of the code implementing this feature.
[1] Sarah Parker, Yue Chen, David Barker, Peter de Rivaz, Debargha Mukherjee, “Global and Locally Adaptive Warped Motion Compensation in Video Compression,” International Conference on Image Processing, pp. 275-279, 2017.
[2] Peter de Rivaz and Jack Haughton, “AV1 Bitstream & Decoding Process Specification”, 2019