BR102017005367A2 - encoding and decoding process of digital images, documents and video files - Google Patents

Abstract

encoding and decoding process of digital image, document and video files new process of the associated process of encoding and decoding digital image, document and video files (coder and decoder), which is the result of studies and developments in response to a Growing demand for the storage and transmission of data, images, documents and videos generated and used in applications, software, equipment, and consumer electronics products used in data centers, cloud storage, medical imaging, remote sensing, banking applications. and notary offices, as well as for application to new devices, equipment and software that require higher storage capacity, transmission capacity, visual quality and speed.

Description

(54) Title: PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS (51) Int. Cl .: H03M 13/00 (73) Holder (s): ANA LUCIA MENDES CRUZ SILVESTRE DA SILVA, FERNANDO SILVESTRE DA SILVA (72) Inventor (s): ANA LUCIA MENDES CRUZ SILVESTRE DA SILVA; FERNANDO SILVESTRE DA SILVA (85) Start date of the National Phase:

03/16/2017 (57) Abstract: PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS New process of the associated device for encoding and decoding digital images, documents and videos (CODER AND DECODER), which is the result of studies and developments in response to a growing demand for the storage and transmission of data, images, documents and videos, generated and used in applications, software, equipment, and consumer electronics products in general, used in data centers, storage in cloud, medical images, remote sensing, banking and notary applications, as well as for application in new devices, equipment and software that require greater storage capacity, transmission capacity, visual quality and speed.

1/23

PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS [001] This report deals with the detailed description accompanied by illustrative figures of a new process for the associated device for encoding and decoding digital images, documents and videos (CODER) AND DECODER), which is the result of studies and developments in response to an increasing demand for the storage and transmission of data, images, documents and videos, generated and used in applications, software, equipment, and consumer electronics products in general, used in data centers, cloud storage, medical images, remote sensing, banking and notary applications, as well as for application in new devices, equipment and software that require greater storage capacity, transmission capacity, visual quality and speed.

Field of invention.

[002] The present invention relates to a new process of encoding and decoding digital images, documents and videos that combines time domain motion estimation / compensation, vector quantization, discrete wavelet transforms, partitioned systems of iterative contractive functions , and also transform coding (or internal prediction), bit plane coding, and / or entropic coding (CODER); as well as their respective inverse processes (DECODER). Developed to be applied in: new devices, applications, equipment, consumer electronics products in general and software with large storage capacity, transmission capacity, visual quality and speed, as well as small high-tech products such as cell phones, data centers, data centers, and bank branch terminals, among others.

Purpose of the Invention.

[003] The process of encoding and decoding digital files of images, documents and videos (CODER AND DECODER), object of the present patent, has as main objective to provide better coding and analysis efficiency in terms of rate

Petition 870170017478, of 03/16/2017, p. 6/41

2/23 compression, subjective quality and rate-distortion ratio of digital images, documents and videos.

State of the art.

[004] There are some patent documents that describe encoding and decoding processes, as well as devices for encoding and decoding, but specifically none of these processes fully anticipates and in the same configuration and procedures as described in this patent. Among these documents, the following can be highlighted:

[005] Patent document PI 0605969-4, DIGITAL IMAGES AND VIDEO ENCODING PROCESS USING WAVELETFRACTAL HYBRID TECHNIQUE, where the present invention refers to a digital image encoding process which employs a combination of fractal encoding associated with the transformed wavelet. In this process, the fractal encoder is applied to the image approximation layer decomposed by a wavelet transform, in order to compress the digital images obtained from visual information, and provide a more efficient way to store and transmit said digital images; [006] The document PI 0605969-4 differs from the present patent application, to quote: the fact of performing the motion estimation in the components of the transformed domain, clearly operating as MRME (motion estimation in multi-resolution); while the present patent application performs motion estimation in the time domain. Similar to components in frequency, components in the transformed domain have completely different nature and characteristics than components in the time domain. In this document PI 0605969-4 only fractal compression is performed in the low-pass sub-images; while the present patent application performs a combination of iterative contractive function systems with vector quantization, with internal prediction and least squares methods; with variation of coding by transform with option on the part of the user to use Haar transform, Hadamard transform, DCT transform or variants. In PI 0605969-4 pure Wavelet transform is used and restricted to three or five levels of decomposition; while the present patent application, in addition to using wavelets differently from PI 0605969-4, uses the

Petition 870170017478, of 03/16/2017, p. 7/41

3/23 image partitioning and data overlapping techniques; and additionally it does not restrict decomposition levels.

[007] The patent document PI 0605969-4 uses SPIHT algorithm on the image as a whole, while the present patent application applies the SPIHT algorithm on the partitioned image; patent document PI 0605969-4 also uses binary coding, whereas the present patent application uses adaptive coding, both of which are considerably different in both performance and complexity.

[008] The patent document C1 0605969-4, DIGITAL IMAGES AND VIDEO ENCODING PROCESS USING THE WAVELETFRACTAL HYBRID TECHNIQUE, where this certificate of addition refers to a process of encoding digital images and videos using the wavelet-fractal hybrid technique . The objective of the present coding process is to compress visual information digitally in order to provide a more efficient way to store and transmit digital images and videos. The main advantages of this process are the elimination of the block effect present in block-based encoders, the high potential for parallelism of the fractal part, the competitive performance curves with mpeg-2, the scalability inherent to compression by wavelet transform, reduction the characteristic blurring effect of the wavelet transform and the high optimization potential yet to be explored;

[009] Patent document PI 112013028615-6 (PCT EP2012 / 057344, EP 11 165099.0), METHOD FOR CODING A SERIES OF SCANNED IMAGES, METHOD FOR DECODING A SERIES OF SCANNED IMAGES, METHOD OF METHODS OF SCANNED IMAGES, DEVICE FOR DECODING A SERIES OF SCANNED IMAGES, ENCODER FOR ENCODING AND DECODING A SERIES OF SCANNED IMAGES, where the present invention relates to a method for encoding a sequence of scanned images (i) that consist of a plurality of pixels. For each of the images, a signal (s) is encoded depending on the image content of the images in question, the non-coded signal being reconstructed and the images

Petition 870170017478, of 03/16/2017, p. 8/41

4/23 reconstructed and derived from it in the course of the coding process. Said reconstructed images (ri) undergo filtration (If) in which a particular reconstructed image (ri) is divided into divisions (pa1, pa2) with one or more filter parameters (fp) being defined for each division (pa1, pa2 ). According to the invention, at least some of the divisions (pi1, pi2) are respectively described using one or more parameters of a function (f1, f2, f3) that describes the pixel curve within a predetermined image region (b) , said pixel curve dividing the region of the predetermined image into at least two divisions (pa1, pa2);

[010] The patent document PI 0408835-2 (PCT CA2004 / 000464, CA 2,423,618), METHOD AND APPARATUS FOR CODE OF RESIDUAL MOVEMENT FRAME WITH TRANSFORMED BASE WITH SUPERCOMPLETE BASE FOR VIDEO COMPRESSION, where the present invention refers to a method for compressing mobile digital images or video signals based on a super-complete base transform using a modified matching search algorithm. More specifically, this invention relates to the efficient coding of the residual motion image, which is generated by the motion estimation and compensation process. A residual energy segmentation algorithm (resa) can be used to obtain an initial estimate of the shape and position of high energy regions in the residual image. A phase-out algorithm (pea) can be used to reduce the number of matching evaluations in the matching search process. The resa or pea can accelerate the encoder many times to find the matched base from the pre-specified supercomplete base dictionary. Three parameters of the matched pattern form an atom, which defines the index in the dictionary and the position of the selected base, as well as the internal product between the chosen base pattern and the residual sign. The present invention provides a new method of atom position encoding using techniques such as quadruple tree and a new atom module quantification scheme. A simple and efficient adaptive mechanism is provided for the design of quantification and position coding to allow a system according to the present invention to operate properly in low, medium and high bit rate situations. These new algorithm components can

Petition 870170017478, of 03/16/2017, p. 9/41

5/23 result in a faster encoding process and improved compression performance compared to video encoders based on previous match searches.

[011] Patent document PI 0904534-1 (PCT JP 2008-300342, JP 2009-089678), METHOD FOR DECODING VIDEOS, which comprises the steps to perform, in relation to an objective decoding area (501) of the objective frame decoding (500 502), a motion search using images from several frames finished with decoding and deciding, based on the result of the motion search, whether an image of the objective decoding area (501) should be generated through a process of interpolation or a decoded image must be generated through motion compensation using data included in an encoding stream.

[012] Patent document PI 9508282-4 (PCT US 9508823, US 275945), METHOD OF UNDERSTANDING DIGITAL DATA, METHOD OF UNDERSTANDING AND RECONSTRUCTING DIGITAL DATA, METHOD OF TRANSMITTING A DIGITAL COMPRESSED DATA COMPRESSED DATA SIGNAL IMAGE UNDERSTANDING SYSTEM, which describes an effective, computationally fast method for compressing digital images using an optimized subsampling process. A set of key functions with spatial overlap (1014) is used as a basis on which the image data (1012) is projected using a mean least squares criterion. Specific processes can be implemented for compression and interpolation of n-dimensional digital data arrangements. Linear, flat and hyperplane key functions allow for optimal, convenient, fast and efficient compression in a closed manner, a process that is easily incorporated into existing digital processing systems. A key advantage of the disclosed method is the fast coding / reconstruction speed, because it involves only convolution or fft types of processors.

[013] Patent document PI 9506690-0 (PCT GB 9424436.5, GB 9424436.5), DIGITAL VIDEO SIGNALS ENCODING PROCESS, DIGITAL VIDEO SIGNALS ENCODER, IMAGE IMAGE SEGMENT

Petition 870170017478, of 03/16/2017, p. 10/41

6/23

DIGITAL VIDEO ENCODED BY THE PROCESS AND OPTICAL DISK, here a process and device are presented for encoding video signals in the form of video image segments (a, b) to enable them to be connected continuously without requiring the restoration of a decoder to a starting state. The system uses an encoder having an encoding stage and an encoding buffer, and successively comprises encoding the images of a video image segment (clip) according to a predetermined encoding scheme (conveniently according to mpeg standards), making the reading-writing of the encoded images in the buffer, and subsequently reading-emitting the video image segment (clip) from the buffer at a substantially constant bit rate. To enable simple segment joining, a predetermined encoding buffer occupation (b ~ 1 c ~) is specified with a controllably varied number of object bits being used to encode an image. Object determination produces an encoding buffer occupation substantially equal to the predetermined buffer occupation (b ~ ic ~) at the time the last segment image has been read-written to the buffer. More specifically for the technique resides in interactive video systems where the user can affect the flow of a narrative without having discontinuity in the presentation of that narrative;

[014] The patent document PI 9800848-0 (PCT US 60 / 040.120, US 60 / 042.245, US 08 / 944.118), PROGNOSIS AND CODIFICATION OF BI-DIRECTIONALLY PROGNOSTICATED VIDEO OBJECT PLANS FOR DIGITAL INTERLATED VIDEO, which describes a system for encoding digital video images such as bi-directionally predicted video object planes (bvops), in particular, where the b-vop and / or the reference image used to encode the b-vop is encoded interlaced. For a b-vop macroblock that is co-located with a predicted field macroblock from a future anchor image, the direct prognosis mode is done by calculating four field motion vectors, then generating the prognostic macroblock. The four field motion vectors and their reference fields are determined from (1) a displacement term of the current macroblock coding vector, (2) of the two field motion vectors of the future anchor image, (3) of the reference field used by the two vectors

Petition 870170017478, of 03/16/2017, p. 11/41

7/23 movement of the field of the co-located future anchor macroblock, and (4) of the time spacing, in the periods of the field, between the current b-vop fields and the anchor fields. Additionally, a coding mode decision process for the current mb selects a forward, backward or average field coding mode, according to a minimum sum of the absolute difference error (sad) that is obtained over the fields top and bottom of the current mb;

[015] Patent document PI 0304545-5 (PCT IB03 / 00061, US 60 / 348,811), IMAGE ENCODING METHOD IN A DIGITAL VIDEO SEQUENCE TO PROVIDE ENCODED VIDEO DATA, VIDEO ENCODER, METHOD OF ENCODING THE INDICATIVE DATA OF A DIGITAL VIDEO SEQUENCE, VIDEO DECODER, AND VIDEO ENCODING SYSTEM, which describes a video encoding system for encoding a video sequence including a sequence of video frames, each having a value chain pixel, where the interpolation filter is set to reconstruct the pixel values in the decoding process. The system encodes the interpolation filter coefficients differently with respect to the predefined base filter, to provide a group of difference values. As the coefficients of the base filter are known to both the encoder and the decoder, these can be reasonably and statically close to the current filters used in the video stream, the decoder can then reconstruct the pixel values based on the group of difference values; and [016] Patent document PI 112013032956-4 (PCT FR2012 / 051391, FR 1155606), IMAGE ENCODING AND DECODING PROCESS, ENCODING AND DECODING DEVICE, AND CORRESPONDING COMPUTER PROGRAMS, where the present invention refers to an encoding process, comprising the clipping (c1) of the image in a plurality of blocks (mb) capable of containing symbols that belong to a predetermined set of symbols, the grouping (c2) of blocks in a predetermined number (p) of block subsets (se1, se2, ... sek, ..., sep), the encoding (c3) by means of an entropic coding module, of each of these block subsets, by associating numerical information with the symbols of

Petition 870170017478, of 03/16/2017, p. 12/41

8/23 each block of a subset considered, this coding step comprising, for the first block of the image, a subset (c33) of initialization of state variables of this entropic coding module, then the generation of a data subflow representative of at least one of these subsets of coded blocks. In the case where the current block is the first block to be coded for a considered subset, determining the probabilities of the appearance of a symbol for that first current block, those probabilities are those that were determined for a predetermined block encoded and decoded from at least one other. subset. In the case where the current block is the last coded block of the considered subset, the writing (c45) proceeds, in the representative subflow of at least one considered subset, of the totality of the numerical information that was associated with the symbols, when coding the data. blocks of that considered subset, and the initialization substep (c46) is used.

Understanding Compression Standards:

[017] Various compression standards can be found in the literature and in practice, such as: JPEG 2000, JPEGXR, Google WebP, MPEG2, ο H264, and ο H265 that stand out for the excellent results in image encoding and compression static and videos.

[018] In image compression standards, the image is treated as a frame (intra-frame), since it is only subjected to intraframe compression. In video standards, frames can be intra-frames, when they are only subjected to intraframe compression or inter-frames, when they are submitted to motion estimation, interframe.

[019] The core of these systems lies in two aspects: in the first aspect for intra-frames with transform compression, for example, wavelet transformed to JPEG2000, transformed DCT to MPEG-2, and modified DCT transformed to H264 and H265; and in the second aspect exclusively for inter-frames, in motion estimation, for example, in the time domain for MPEG-2, H264, H265.

Petition 870170017478, of 03/16/2017, p. 13/41

9/23 [020] In sequence to transform coding, quantization is usually applied, such as: linear quantization, vector quantization or bit plane coding, and also entropic coding, such as: Huffman and arithmetic coding. Inverse processes similar to encoding take place in decoding, however, performed in the inverse sequence, usually through the mirrored process, but not necessarily.

[021] It should be noted that the standards mentioned above do not use partitioned systems with iterative contractive functions in their composition.

Understanding Vector Quantization Compression, Wavelets, Iterative Contractive Partitioned Systems, Bit Plane Coding, Entropic Coding [022] The principle of transform coding is the application of linear transformations in a signal, producing a more efficient representation of it in a different domain. This domain change allows subsequent processes to be applied advantageously in a way never possible in the original domain.

[023] The application of wavelets in an image results in the decomposition of this same image in 4 (four) new sub-images at the level of decomposition 1 (one), which results in a low-pass sub-image, a sub-image of vertical details, a sub-image of details horizontal and a subimage of diagonal details. This procedure is applied recursively in the low-pass subimage generated for the other levels, generating a total of 3k + 1 sub-images, where k is the decomposition level. In some cases, this transform can be implemented via fast algorithms to reduce computational complexity. Among the advantages of Wavelets, we can mention the fact of having: fast algorithms, excellent concentration of energy in a few coefficients, progressive transmission, balance between encoding and decoding speed and the absence of blocking. Some artifacts and distortions such as ringing, blurring, among others, however, remain a challenge to be overcome.

[024] The partitioned systems of iterative contractive functions explore the principle of self-similarity between blocks of different sizes, shapes and

Petition 870170017478, of 03/16/2017, p. 14/41

10/23 orientations within the image itself, establishing the relationship between them through similar transformations. The few coefficients of these transformations are transmitted, instead of the pixels of the image, resulting in significant bit savings. The image is originally partitioned into blocks called “range-blocks”, which are compared with the “domain-blocks” blocks, which are larger in size than the “range-block” blocks in question, within a “domain- pool ”(domain area) through search processes for the best“ matching ”. Once the related transformation that maps each range-block, that is, the best matching, is determined, only the coefficients are transmitted. Self-similarity comes from the fact that the "domain pool" is formed by blocks that belong to the same region where the "range-block" comes from. Among the advantages of this compression, we can mention: the high quality, the recovery of the details of the image, the extensive saving of bits for storage / transmission, and the decoding speed. Among the disadvantages are blocking artifacts and the extreme computational weight of coding.

[025] Vector quantization consists of representing blocks of data by a linear combination of code words selected from a pre-established code book, increasing the efficiency of coding. A poorly constructed codebook can cause significant and irreparable distortions in the reconstructed data.

[026] Bit-plane coding consists of analyzing a set of data organized not according to its spatial location, but rather by a classification related to its amplitude in relation to a dyadic threshold. With its use it is possible to hierarchically explore the redundancies of the image representation, leading to a more compact representation of the data.

[027] Entropic coding aims at compressing the representation of digital data by reducing the possible statistical redundancy between them. Such redundancy is evidenced by the measure of Entropy within the Information Theory. [028] Adaptive arithmetic coding is an entropic coding method that uses dynamic ranges that are selected, subdivided and scaled adaptively according to the statistical distribution of the signal being processed.

Petition 870170017478, of 03/16/2017, p. 15/41

11/23

Understanding video compression as motion estimation / compression.

[029] Video compression techniques, generally, emerge with the objective of exploring the temporal redundancy contained between frames, directly sequential or not, of a video sequence (movements), where the processing is performed block by block of pixels.

[030] The Motion Estimation of a given block in the current frame can be performed through numerous search and scan techniques applied by comparing the current frame with all or a subset of the pixel blocks of another frame, called the reference frame. By minimizing a cost function, the equivalent block is determined in the reference frame, which most closely resembles the current block (in the current frame). Only the difference between both frames / blocks (signal-error or residual-frame) resulting from Motion Compensation, will pass to subsequent processing steps. Therefore, only the difference between frames is transmitted / stored. In Motion Estimation, several search techniques can be adopted in place of the exhaustive search that considers absolutely all possibilities of block-equivalent.

Description of the figures.

[031] Below, reference is made to the Figures accompanying this descriptive report, for a better understanding and illustration of it, where it can be seen:

Figure 1 shows a diagram of the block diagram of the Process Encoder and device for encoding and decoding digital images, documents and videos, object of the present patent.

Figure 2 shows a scheme of the block diagram of the Process Decoder and device for encoding and decoding digital images, documents and videos, object of the present patent.

Description of the invention.

[032] In the process of encoding and decoding digital files of images, documents and videos (CODER AND DECODER), the object of this patent, encoding is a process for encoding digital images, documents and videos that combines estimation / motion compensation in the time domain,

Petition 870170017478, of 03/16/2017, p. 16/41

12/23 vector quantization, discrete wavelet transforms, iterative contractive function systems, bit plane coding, internal prediction or transform coding, and entropic coding, where the coding comprises eight steps in sequence of:

- a) Estimation / Compensation of Motion in the time domain, performed in the inter frames of the video sequence, generating residual frames;

- b) Application of a discrete wavelet transform, with the option of partitioning by the user, including whether or not using overlap implemented through lifting techniques applied to the image frames, in the following three situations: in the first situation to the intra-frames of the video sequence; in the second situation to the residual frames of the motion estimation / compensation performed in the video sequence; and in the third situation to the intra-frame that composes the static image.

- c) Application of a combination of vector quantization with coding by systems of iterative contractive functions, in the sub-images, in the following two situations: in the first situation to the low-pass sub-images resulting from the application of the discrete wavelet transform with partitioning option, obtained from previous step, using or not overlapping the intra frames of the video sequence; and in the second situation, the low-pass subimage resulting from the application of the wavelet transform with partitioning option, using or not overlapping the intra-frame that composes the static image, obtained from the previous step; being that in the two situations above for iterative contractive function systems, the search process for the best “matching” is performed only within the same subimage that is being processed.

- d) Application of bit plane encoding to the residual frames of the motion sequence estimation / compensation of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; for this, a hierarchical algorithm is used to support partitioning and overlap.

Petition 870170017478, of 03/16/2017, p. 17/41

13/23

- e) Application of bit plane encoding to high-pass sub-images belonging to: 1) intra-frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap, 2) as well as sub-images walkways belonging to the intra-frame of the static image that was transformed via a discrete wavelet transform with partitioning option, using or not overlap; being used in both 1 and 2, for this, an adapted hierarchical algorithm that supports partitioning and overlap.

- f) Application of an internal prediction with least squares methods; or applying a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the blocks of the low-pass subimage belonging to the following situations: in the first situation, the intra frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; in the second situation to the intra-frame of the static image that was transformed via a discrete wavelet transform with partitioning option, using or not overlap; being applied in both situations if they do not reach a minimum quality criterion established by the user.

- g) Application of entropic coding to all signals resulting from bit plane coding by adaptive arithmetic coding.

[033] In the process of encoding and decoding digital files of images, documents and videos (CODER AND DECODER), object of the present patent, in the encoding the Motion Estimation / compensation occurs in the time domain, being applied in the inter-frames of the video sequence, generating residual frames that will undergo spatial compression composed of a combination of vector quantization, iterative contractive function systems, bit plane coding, internal prediction or transform coding, and entropic coding. The domain change occurs through the application of a discrete wavelet transform with the option of partitioning by the user, including whether or not

Petition 870170017478, of 03/16/2017, p. 18/41

14/23 overlap implemented through lifting techniques, to the intra-frame that composes the static image, which will undergo spatial compression (composed of a combination of vector quantization, iterative contractive function systems, bit plane coding, internal prediction or transform coding, and entropic coding); and the domain change occurs through the application of a discrete wavelet transform with the option of partitioning by the user, including whether or not using overlap implemented through lifting techniques, to intra-frames and residual frames of motion estimation / compensation of the video sequence, both of which will undergo spatial compression composed of a combination of vector quantization, iterative contractive function systems, bit plane encoding, internal prediction or transform encoding, and entropic encoding.

[034] In the coding process, the application of a combination of vector quantization and coding by systems of iterative contractive functions is used, in the following situations: in the first situation, in the low-pass sub-images resulting from the application of the wavelet transform with partitioning option, using or not overlapping the intra frames of the video sequence; in the second situation in the low-pass subimage resulting from the application of the discrete wavelet transform with partitioning option, using or not overlap, of the inside frame that composes the static image; and in both situations in the iterative contractive function systems, the search is performed only within the same subimage that is being processed.

[035] In the coding process, a classification method with restrictions on the minimum number of blocks is used to accelerate the search procedure of the iterative contractive function system.

[036] In the coding process the application of bit plane encoding is performed in the following situations: in the first situation, a hierarchical algorithm that supports partitioning and overlap, in the residual frames of the motion estimation / compensation of the video sequence that were transformed via a discrete wavelet transform with partitioning option, using or not overlap; and in the second situation hierarchical algorithm that supports the

Petition 870170017478, of 03/16/2017, p. 19/41

15/23 partitioning and overlap, in the high-pass sub-images belonging to the inside frames of the video sequence, as well as in the high-pass sub-images belonging to the intra-frame of the static image; both situations being transformed via discrete wavelet transform with partitioning option, using or not overlap. [037] In the coding process, an internal prediction with least squares methods is applied; or applying a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the blocks of the low-pass subimage belonging to the following situations: in the first situation, the intra-frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; and in the second situation to the intra-frame of the static image that was transformed via discrete wavelet transform with partitioning option, using or not overlap; being applied in both situations if they do not reach a minimum quality criterion established by the user.

[038] In the encoding process, entropic encoding is applied to all signals resulting from bit plane encoding, by adaptive arithmetic encoding. [039] In the process of encoding and decoding digital images, documents and videos (CODER AND DECODER), object of this patent, decoding is a process of decoding digital images, documents and videos that combines entropic decoding , internal prediction or transform encoding, bit plane decoding, iterative contractive function systems, inverse discrete wavelet transform, inverse vector quantization, and time domain motion compensation; where decoding comprises eight steps in reverse encoding sequence, as follows:

- a) Application of entropic decoding to the bitstream resulting from encoding by adaptive arithmetic decoding.

- b) Application of an internal prediction with least squares methods; or applying a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out

Petition 870170017478, of 03/16/2017, p. 20/41

16/23 only in the part of the bitstream referring to the blocks of the passimage subimage belonging to two situations: in the first situation to the intra frames of the video sequence; and in the second situation to the intra-frame of the static image.

- c) Application of bit plane decoding to the portion of the bitstream that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, to form the residual frames to undergo motion compensation of the video sequence; being used, for decoding by bit plane, inverse hierarchical algorithm that supports partitioning and overlap.

- d) Application of decoding by bit planes to the portion of the bitstream that will form the high-pass sub-images that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, to form: the intra frames of the video sequence; and the intra-frame of the static image; being used for decoding by bit plane, a hierarchical inverse adapted algorithm that supports partitioning and overlap.

- e) Application of a combination of inverse vector quantization with decoding by systems of iterative contractive functions to the data that will form the low-pass sub-images that will undergo inverse discrete wavelet transformation with partitioning option, using or not overlap, to compose: intra of the video sequence; and the intra-frame of the still image.

- f) Application of a discrete inverse wavelet transform, with partitioning option, using or not overlap, implemented through lifting techniques, to the images that will form: the intra frames of the reconstructed video sequence; the residual frames that will undergo motion compensation to reconstruct the video sequence, and the intra frame of the reconstructed static image.

- g) Compensation of Movement in the time domain, performed in the residual frames that will form the inter-frames of the reconstructed video sequence.

Petition 870170017478, of 03/16/2017, p. 21/41

17/23 [040] In the decoding process, entropic decoding is applied to the bitstream resulting from the encoding, by adaptive arithmetic decoding. An internal prediction with least squares methods is also applied; or application of a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the part of the bitstream referring to the blocks of the low-pass sub-image belonging to: the intra frames of the video sequence; and the intra-frame of the static image.

[041] In the decoding process, decoding by bit plane is applied, being: inverse hierarchical algorithm that supports partitioning and overlap, to the portion of the bitstream that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, to form the residual frames for the motion compensation of the video sequence; and inverse adapted hierarchical algorithm that supports partitioning and overlap, to the bitstream portion that will form the high-pass sub-images that will undergo an inverse discrete wavelet transform with partitioning option, using or not overlap, to compose the intra frames of the video sequence , as well as the bitstream portion that will form the high-pass sub-images that will undergo an inverse discrete wavelet transform with partitioning option, using or not overlap, to compose the intra-frame of the static image.

[042] In the decoding process, a combination of inverse vector quantization and decoding by systems of iterative contractive functions is also applied to the data that will form the low-pass sub-images that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, for compose: the intra frames of the video sequence; and the intra-frame of the still image.

[043] In the decoding process, the domain change occurs through the application of a discrete inverse wavelet transform, with partitioning option, using or not overlap, implemented through lifting techniques to the data that will form: the intra frames of the sequence of reconstructed video; residual frames for motion compensation in the video stream to be

Petition 870170017478, of 03/16/2017, p. 22/41

Rebuilt 18/23; and the intra-frame of the static image to be reconstructed. Motion Compensation occurs in the time domain, to be applied to the residual frames that will form the inter frames of the reconstructed video sequence.

Preferred way of carrying out the invention [044] The process and device for encoding and decoding digital images, documents and videos (CODER AND DECODER), object of the present patent, refers to a coding process (Figure 1) and decoding process (Figure

2) digital images, documents and videos that combine time domain motion estimation / compensation, vector quantization, discrete wavelet transforms, iterative contractive function systems, bit plane encoding, internal prediction or transform encoding, entropic encoding; as well as their respective inverse processes, which provides better coding efficiency in terms of compression rate, subjective quality and taxadistance ratio.

[045] The aforementioned objective is achieved given the balance obtained between the advantages and disadvantages characteristic of each of the techniques in isolation, when combined. This balance is obtained mainly since the characteristic artifacts are reduced, such as: the ringing, the blocking of the partitioned systems of iterative contractive functions, and the blocking of the estimation / compensation of movement in the time domain; occurring the balance of complexity of the system as a whole between encoder and decoder (non-existent for contractive iterative systems); where positive forces are highlighted as: the speed of the estimation of motion in the time domain, the wavelet speed, the excellence in the coding of details of iterative contractive partitioned systems, and the adaptability to the coding of symbols from the statistical point of view of vector quantization.

[046] In this context the present invention includes a physical and virtual encoder and decoder device, each divided into six fundamental parts, as shown in Figures 1 (encoder) and 2 (decoder).

[047] According to Figure 1, for video coding (102), the first stage consists of temporal compression through the application of the estimation / compensation of movement in the time domain (105), performed in the inter-frames (104 ) gives

Petition 870170017478, of 03/16/2017, p. 23/41

19/23 video sequence, generating residual frames (106), also known as error frames.

[048] Subsequently, they undergo domain transformation when subjected to a discrete wavelet transformation (107), with the option of partitioning by the user, including using or not overlap, implemented through lifting techniques: for image coding and documents, the intra-frame that composes the static image (101); for video coding, the intra frames of the video sequence (103), as well as the residual frames (106) as a result of the motion compensation / estimation performed in the video sequence.

[049] Next, a combination of vector quantization with coding by iterative contractive function systems (110) is applied to the following signals: for image coding, documents, in the low-pass sub-images (108) that resulted from the application of discrete wavelet transform (107) with partitioning option, using or not overlapping the intra-frame that composes the static image (101) to be compressed; for video encoding, in the sub-images passabras (108) that were the result of the application of the discrete wavelet transform (107) with partitioning option, using or not overlapping the intra frames (103) of the video sequence. Since, in both cases (image, documents and video), in these iterative contractive function systems (110), the search process for the best match between blocks is performed only within the same subimage that is being processed.

[050] Additionally, for the iterative contractive function systems (110), the classification method with minimum number of blocks restrictions is used to accelerate the search procedure of the iterative contractive function system, such as the Fisher classification method.

[051] In the next step, bit plane encoding (111 and 112) is applied as follows: For image and document encoding, in the detail sub-images (109) belonging to the still frame of the still image (101) transformed via discrete wavelet transform with partitioning option, using or not overlap (107), and the adapted hierarchical algorithm that supports partitioning and overlap can be preferably a modification of the algorithm

Petition 870170017478, of 03/16/2017, p. 24/41

20/23

SPIHT (Set Partitioned in Hierarchical Trees) (111); For video encoding, in the following situations: in the first situation, in the residual frames (106) of the motion estimation / compensation (105) of the video sequence that were transformed via discrete wavelet transform with partitioning option (107), using or not overlap, and the hierarchical algorithm that supports partitioning and overlap, may preferably be the original SPIHT algorithm (Set Partitioned in Hierarchical Trees) (112); and in the second situation in the high-pass sub-images (109) belonging to the intra-frames of the video sequence (103) that were transformed via discrete wavelet transform with partitioning option (107), using or not overlap, with the hierarchical algorithm adapted that supports partitioning and overlap can preferably be a modification of the SPIHT (Set Partitioned in Hierarchical Trees) algorithm.

[052] In the next step, an internal prediction with least squares methods is applied (113); or application of a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out as follows: For the encoding of images and documents, in the blocks of the low-pass sub-image (108) belonging to the intra-frame of the static image (101) which was transformed via a discrete wavelet transform with partitioning option, using or not overlap (107); and For video encoding, in the blocks of the low-pass sub-image (108) belonging to the intra frames of the video sequence (103) that were transformed via discrete wavelet transform with partitioning option, using or not overlap (107). In both cases, applied if they do not reach a minimum quality criterion established by the user.

[053] Finally, entropy coding (114) is applied to signals resulting from bit plane encodings, by adaptive binary arithmetic coding, generating the encoded bitstream (115).

[054] According to Figure 2, the first stage of decoding consists of applying entropic decoding (202) to the bitstream (201) which is the result of

Petition 870170017478, of 03/16/2017, p. 25/41

21/23 image or video encoding (Figure 1), by adaptive binary arithmetic decoding.

[055] In the next step, an internal prediction with least squares methods is applied (203); or application of a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out as follows: For image and document decoding, only in the part of the bitstream referring to the blocks of the low-pass sub-image (207) belonging to the intra-frame of the static image (214); and for video decoding. only in the part of the bitstream referring to the blocks of the low-pass sub-image (207) belonging to the intra frame of the video sequence (213).

[056] In the next step, decoding by bit plane (205 and 206) is applied, as follows: For decoding images and documents, using the inverse hierarchical algorithm that supports partitioning and overlap, being preferably a modification of the SPIHT (Set Partitioned in Hierarchical Trees) algorithm, (205) in the portion of the bitstream that will form the high-pass sub-images (208) that will undergo an inverse discrete wavelet transform with partitioning option, using or not overlap (209), to compose the frame intra of the static image (214); and for video decoding, through the inverse hierarchical algorithm that supports partitioning and overlap, being preferably the SPIHT (Set Partitioned in Hierarchical Trees) algorithm (206), in the bitstream portion that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap (209), to form the residual frames (210) for motion compensation of the video sequence (211), as well as through a modification of the inverse SPIHT (205) in the bitstream portion that will form the sub-images high-pass (208) that will undergo an inverse discrete wavelet transform with partitioning option, using or not overlap, (209) to form the frames in the video sequence (213).

[057] Then, a combination of inverse vector quantization with decoding by iterative contractive function systems (204) is then applied to the following signals: For image decoding, in the data that will form the subPicture passPetition 870170017478, of 03/16/2017, p. 26/41

22/23 casualties (207) that will undergo a discrete reverse wavelet transformation with partitioning option, using or not overlap (209), to compose the intra-frame of the static image (214); and for video decoding, in the data that will form the low-pass sub-images (207) that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap (209), to compose the intra frames of the video sequence (213) .

[058] Then, the following data undergoes domain transformation when subjected to a discrete reverse wavelet transformation with partitioning option, using or not overlap (209), implemented through lifting techniques: For image decoding, the data that will form the intra-frame that composes the reconstructed static image (214); and for video decoding, the data that will form the intra frames (213) of the reconstructed video sequence (215), as well as the data that will form the residual frames (210) that will undergo motion compensation (211) in the sequence of video.

[059] Finally, movement compensation is performed in the time domain (211), in the residual frames (210) generating the inter frames (212) that together with the intra frames (213) form the sequence of reconstructed video (215).

[060] The present invention is described here in order to enable a clearer understanding of the object in question and its implementation, however, such description does not intend to limit the invention to the form disclosed here. In this context, modifications compatible with the teachings described herein, and the skill or knowledge of the relevant technique, are within the scope of the present invention. It is intended that the present invention includes all modifications and variations thereof, within the scope described in the report and the attached claims.

[061] In this way, the process of encoding and decoding digital files of images, documents and videos (CODER AND DECODER), object of this patent, as described above, presents a new and unique configuration and functionality that give it great advantages in terms of processes of the same nature currently used and available on the market. Among these advantages we can mention: the fact of improving the efficiency of encoding and decoding in terms of compression rate, subjective quality and rate-distortion ratio of

Petition 870170017478, of 03/16/2017, p. 27/41

23/23 digital images, documents and videos; and the fact of uniting and integrating the encoding and decoding of images, documents and videos in one process.

[062] Therefore, due to the characteristics of configuration and operation, described above, it can be clearly noted that the PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, this is a new procedure for the State of the Art o which has unprecedented conditions of innovation, inventive activity and industrialization, which make it deserve the Patent Privilege of Invention.

Petition 870170017478, of 03/16/2017, p. 28/41

1/8

Claims (16)

1 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Process of encoding images, documents and digital videos that combines estimation / compensation of movement in the time domain, vector quantization, discrete wavelet transforms, systems of contractive functions iterative, bit plane coding, internal prediction or transform coding, entropic coding, characterized by the coding comprising eight steps in sequence of: a) Estimation / Compensation of Motion in the time domain, performed in the inter frames of the video sequence, generating residual frames; b) application of a discrete wavelet transform, with the option of partitioning by the user, including whether or not using overlap implemented through lifting techniques applied to image frames, in the following three situations: in the first situation to the intra frames of the sequence of video; in the second situation to the residual frames of the motion estimation / compensation performed in the video sequence; and in the third situation to the internal frame that composes the static image; c) application of a combination of vector quantization with coding by systems of iterative contractive functions, in the sub-images, in the following two situations: in the first situation to the low-pass sub-images resulting from the application of the discrete wavelet transform with partitioning option, obtained from the step above, using or not overlapping the intra frames of the video sequence; and in the second situation, the low-pass subimage resulting from the application of the wavelet transform with partitioning option, using or not overlapping the intra-frame that composes the static image, obtained from the previous step; being that in the two situations above for iterative contractive function systems, the search process for the best “matching” is performed only within the same subimage that is being processed; d) Application of bit plane encoding to the residual frames of the motion estimation / compensation of the video sequence that were transformed via discrete wavelet transform with partitioning option, Petition 870170058413, of 08/14/2017, p. 5/13 2/8 using or not overlap; being used, for this, hierarchical algorithm that supports partitioning and overlap; e) application of bit plane encoding to the high-pass sub-images belonging to: 1) intra-frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap, 2) as well as the sub-images passes -highs belonging to the intra-frame of the static image that was transformed via discrete wavelet transform with partitioning option, using or not overlap; being used in both 1 and 2, for this, an adapted hierarchical algorithm that supports partitioning and overlap; f) application of an internal prediction with least squares methods; or applying a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the blocks of the low-pass subimage belonging to the following situations: in the first situation, the intra frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; in the second situation to the intra-frame of the static image that was transformed via a discrete wavelet transform with partitioning option, using or not overlap; being applied in both situations if they do not reach a minimum quality criterion established by the user; g) Application of entropic coding to all signals resulting from bit plane coding by adaptive arithmetic coding. 2 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Coding process according to claim 1, characterized by the motion estimation / compensation occurring in the time domain, and being applied in the inter frames of the sequence of video, generating residual frames that will undergo spatial compression composed of a combination of vector quantization, iterative contractive function systems, bit plane encoding, internal prediction or transform encoding, and entropic encoding. Petition 870170058413, of 08/14/2017, p. 6/13 3/8 3 - ENCODING AND DECODING PROCESS OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Encoding process according to claim 1, characterized by the domain change occurring through the application of a discrete wavelet transform with the option of partitioning by the user, whether or not using overlap implemented through lifting techniques, the internal frame that composes the static image, which will undergo spatial compression composed of a combination of vector quantization, iterative contractive function systems, bit plane coding, internal prediction or transform coding , and entropic coding. 4 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Coding process according to claim 1, characterized by the domain change occurring through the application of a discrete wavelet transform with the option of partitioning by the user, whether or not using overlap implemented through lifting techniques, in the intra frames and in the residual frames of the video sequence motion estimation / compensation, both of which will undergo spatial compression composed of a combination of vector quantization, systems of contractive functions iterative, bit plane coding, internal prediction or transform coding, and entropic coding. 5 - PROCESS OF ENCODING AND DECODING OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Coding process according to claim 1, characterized by the application of a combination of vector quantization with coding by systems of iterative contractive functions to be carried out in the following situations: in the first situation in the low-pass sub-images resulting from the application of the wavelet transform with partitioning option, using or not overlap, of the intra frames of the video sequence; and in the second situation in the low-pass subimage resulting from the application of the discrete wavelet transform with partitioning option, using or not overlap, of the intra-frame that composes the static image; in both situations, in iterative contractive function systems, the search is performed only within the same subimage that is being processed. Petition 870170058413, of 08/14/2017, p. 7/13 4/8 6 - ENCODING AND DECODING PROCESS OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Encoding process according to claims 1 and 5, characterized by using a classification method with restrictions on the minimum number of blocks to accelerate the system search procedure of iterative contractive functions. 7 - ENCODING AND DECODING PROCESS OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Encoding process according to claim 1, characterized by the application of bit plane encoding to be carried out in the following situations: in the first situation, hierarchical algorithm that supports partitioning and overlap, in the residual frames of motion estimation / compensation of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; and in the second situation, an adapted hierarchical algorithm that supports partitioning and overlap, in the high-pass sub-images belonging to the inside frames of the video sequence, as well as, in the high-pass sub-images belonging to the intra-frame of the static image; both situations having been transformed via a discrete wavelet transform with partitioning option, using or not overlap. 8 - CODING AND DECODING PROCESS OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Coding process according to claims 1 and 5, characterized by applying an internal prediction with least squares methods; or apply a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in blocks of the low-pass subimage belonging to the following situations: in the first situation, the intra frames of the video sequence that were transformed via discrete wavelet transform with partitioning option, using or not overlap; and in the second situation to the intra-frame of the static image that was transformed via discrete wavelet transform with partitioning option, using or not overlap; being applied in both situations, if they do not reach a minimum quality criterion established by the user. Petition 870170058413, of 08/14/2017, p. 8/13 5/8 9 - ENCODING AND DECODING PROCESS OF DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Encoding process according to claims 1 and 7, characterized by applying entropic encoding to all signals resulting from bit plane encoding, by adaptive arithmetic encoding . 10 - ENCODING AND DECODING PROCESS OF DIGITAL IMAGES, DOCUMENTS AND VIDEOS FILES, Digital image, document and video decoding process that combines entropic decoding, internal prediction or transform encoding, bit plane decoding, iterative contractive function systems , inverse discrete wavelet transform, inverse vector quantization, and time domain motion compensation; characterized by the decoding comprising eight steps in reverse encoding sequence, of: a) Application of entropic decoding to the bitstream resulting from encoding by adaptive arithmetic decoding; b) Application of an internal prediction with least squares methods; or applying a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the part of the bitstream referring to the blocks of the low-pass subimage belonging to two situations: in the first situation to the intra frames of the video sequence; and in the second situation to the intra-frame of the static image; c) Application of decoding by bit planes to the bitstream portion that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, to form the residual frames to undergo motion compensation of the video sequence; being used, for decoding by bit plane, inverse hierarchical algorithm that supports partitioning and overlap; d) Application of bit plane decoding to the portion of the bitstream that will form the high-pass sub-images that will undergo reverse discrete wavelet transformation with partitioning option, using or not overlap, to form: Petition 870170058413, of 08/14/2017, p. 9/13 6/8 the intra-frames of the video sequence; and the intra-frame of the static image; being used for decoding by bit plane, a hierarchical inverse adapted algorithm that supports partitioning and overlap; e) Application of a combination of inverse vector quantization with decoding by systems of iterative contractive functions to the data that will form the low-pass sub-images that will undergo inverse discrete wavelet transformation with partitioning option, using or not overlap, to compose: the intra-frames the video stream; and the intra-frame of the static image; f) Application of a discrete inverse wavelet transform, with partitioning option, using or not overlap, implemented through lifting techniques, to the images that will form: the intra frames of the reconstructed video sequence; the residual frames that will undergo motion compensation to reconstruct the video sequence, and the intra frame of the reconstructed static image; g) Compensation of Movement in the time domain, performed in the residual frames that will form the inter-frames of the reconstructed video sequence. 11 - ENCODING AND DECODING PROCESS OF DIGITAL IMAGES, DOCUMENTS AND VIDEOS FILES, Decoding process according to claim 10, characterized by applying entropic decoding to the bitstream resulting from encoding, by adaptive arithmetic decoding. 12 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Decoding process according to claim 10, characterized by the application of an internal prediction with least squares methods; or application of a variation of coding by transform with the user's option to use Haar transform, Hadamard transform, DCT transform or variants; both applications being carried out only in the part of the bitstream referring to the blocks of the low-pass sub-image belonging to: the intra frames of the video sequence; and the intra-frame of the static image. Petition 870170058413, of 08/14/2017, p. 10/13 7/8 13 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Decoding process according to claim 10, characterized by the application of decoding by bit plane, to be: inverse hierarchical algorithm that supports partitioning and overlap, the portion of the bitstream that will undergo an inverse discrete wavelet transformation with the option of partitioning, using or not overlapping, to form the residual frames for the motion compensation of the video sequence; and inverse adapted hierarchical algorithm that supports partitioning and overlap, in the portion of the bitstream that will form the high-pass sub-images that will undergo an inverse discrete wavelet transformation with partitioning option, using or not overlap, to compose the intra frames of the video sequence , as well as in the portion of the bitstream that will form the high-pass sub-images that will undergo an inverse discrete wavelet transform with partitioning option, using or not overlap, to compose the intra-frame of the static image. 14 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Decoding process according to claim 10, characterized by the application of a combination of inverse vector quantization with decoding by systems of iterative contractive functions to the data that iterates. they will form the low-pass sub-images that will undergo a discrete reverse wavelet transformation with partitioning option, using or not overlap, to compose: the intra frames of the video sequence; and the intra-frame of the still image. 15 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Decoding process according to claim 10, characterized by the domain change occurring through the application of an inverse discrete wavelet transform, with option of partitioning, using or not overlap, implemented through lifting techniques, to the data that will form: the intra frames of the reconstructed video sequence; residual frames for motion compensation in the video sequence to be reconstructed; and the intra-frame of the static image to be reconstructed. Petition 870170058413, of 08/14/2017, p. 11/13 8/8 16 - PROCESS OF ENCODING AND DECODING DIGITAL FILES OF IMAGES, DOCUMENTS AND VIDEOS, Decoding process according to claim 10, characterized by the Motion Compensation occurring in the time domain, to be applied in the residual frames that will form the frames inter of the reconstructed video sequence. Petition 870170058413, of 08/14/2017, p. 12/13 1/1 115