Patents by Inventor Roger Kumar
Roger Kumar has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20110188575Abstract: Some embodiments of the invention provide a method for encoding a video signal that is formed by a series of successive images. Each image includes several sections, and each section has a set of image values. To encode a particular section of a particular image, the method initially partitions the particular section into several sub-sections. For each of at least two particular sub-sections, the method then computes a statistical parameter regarding the image values of the particular sub-section. The method compares the computed statistical parameters, and based on the comparison, selects an encoding technique from a set of encoding techniques to encode the particular section. In some embodiments, the set of encoding schemes includes a first scheme that encodes the selected section without reference to any other section of any other image, and a second scheme that encodes the selected section by reference to at least one other section.Type: ApplicationFiled: April 12, 2011Publication date: August 4, 2011Inventors: Thomas Pun, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20110182360Abstract: Some embodiments of the invention provide a method for encoding a video signal that is formed by a series of successive images. Each image includes several sections, and each section has a set of image values. To encode a particular section of a particular image, the method initially partitions the particular section into several sub-sections. For each of at least two particular sub-sections, the method then computes a statistical parameter regarding the image values of the particular sub-section. The method compares the computed statistical parameters, and based on the comparison, selects an encoding technique from a set of encoding techniques to encode the particular section. In some embodiments, the set of encoding schemes includes a first scheme that encodes the selected section without reference to any other section of any other image, and a second scheme that encodes the selected section by reference to at least one other section.Type: ApplicationFiled: April 8, 2011Publication date: July 28, 2011Inventors: Thomas Pun, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7944971Abstract: Some embodiments of the invention provide a method for encoding a video signal that is formed by a series of successive images. Each image includes several sections, and each section has a set of image values. To encode a particular section of a particular image, the method initially partitions the particular section into several sub-sections. For each of at least two particular sub-sections, the method then computes a statistical parameter regarding the image values of the particular sub-section. The method compares the computed statistical parameters, and based on the comparison, selects an encoding technique from a set of encoding techniques to encode the particular section. In some embodiments, the set of encoding schemes includes a first scheme that encodes the selected section without reference to any other section of any other image, and a second scheme that encodes the selected section by reference to at least one other section.Type: GrantFiled: May 1, 2003Date of Patent: May 17, 2011Assignee: Apple Inc.Inventors: Thomas Pun, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20110019879Abstract: Some embodiments of the invention provide a method for adaptively performing motion estimation. In some embodiments, the method initially performs a first motion estimation operation. It then determines whether the estimated error in the result of the first motion estimation is greater than a particular threshold. If not, the method uses the results of the first motion estimation operation. Otherwise, the method selects another motion estimation operation. The method continues in this manner until it performs a motion estimation operation that results in an acceptable estimated error, or until it tries all the motion estimation operations at its disposal. When the method cannot identify a motion estimation operation that produces results with an acceptable estimated error, the method selects the result of the motion estimation operation that produced the best results.Type: ApplicationFiled: May 3, 2010Publication date: January 27, 2011Inventors: Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20110007798Abstract: Method For Implementing A Quantizer In A Multimedia Compression And Encoding System is disclosed. In the Quantizer system of the present invention, several new quantization ideas are disclosed. In one embodiment, adjacent macroblocks are grouped together into macroblock groups. The macroblock groups are then assigned a common quantizer value. The common quantizer value may be selected based upon how the macroblocks are encoded, the type of macroblocks within the macroblock group (intra-blocks or inter-blocks), the history of the motion vectors associated with the macroblocks in the macroblock group, the residuals of the macroblocks in the macroblock group, and the energy of the macroblocks in the macroblock group. The quantizer value may be adjusted in a manner that is dependent on the current quantizer value. Specifically, if the quantizer value is at the low end of the quantizer scale, then only small adjustments are made.Type: ApplicationFiled: June 21, 2010Publication date: January 13, 2011Inventors: Thomas Pun, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7769084Abstract: Method For Implementing A Quantizer In A Multimedia Compression And Encoding System is disclosed. In the Quantizer system of the present invention, several new quantization ideas are disclosed. In one embodiment, adjacent macroblocks are grouped together into macroblock groups. The macroblock groups are then assigned a common quantizer value. The common quantizer value may be selected based upon how the macroblocks are encoded, the type of macroblocks within the macroblock group (intra-blocks or inter-blocks), the history of the motion vectors associated with the macroblocks in the macroblock group, the residuals of the macroblocks in the macroblock group, and the energy of the macroblocks in the macroblock group. The quantizer value may be adjusted in a manner that is dependent on the current quantizer value. Specifically, if the quantizer value is at the low end of the quantizer scale, then only small adjustments are made.Type: GrantFiled: April 30, 2003Date of Patent: August 3, 2010Assignee: Apple Inc.Inventors: Thomas Pun, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7742525Abstract: A method for adaptively performing motion estimation. In some embodiments, the method initially performs a first motion estimation operation. It then determines whether the estimated error in the result of the first motion estimation is greater than a particular threshold. If not, the method uses the results of the first motion estimation operation. Otherwise, the method selects another motion estimation operation. The method continues in this manner until it performs a motion estimation operation that results in an acceptable estimated error, or until it tries all the motion estimation operations at its disposal. When the method cannot identify a motion estimation operation that produces results with an acceptable estimated error, the method selects the result of the motion estimation operation that produced the best results.Type: GrantFiled: April 30, 2003Date of Patent: June 22, 2010Assignee: Apple Inc.Inventors: Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20090010325Abstract: A rate controller for allocating a bit budget for video frames to be encoded is disclosed. The rate controller of the present invention considers many different factors when determining the frame bit budget including: desired video quality, target bit rate, frame type (intra-frame or inter-frame), frame duration, intra-frame frequency, frame complexity, intra-block frequency within an intra-frame, buffer overflow, buffer underflow, and the encoded video frame quality for a possible second pass.Type: ApplicationFiled: June 30, 2008Publication date: January 8, 2009Inventors: Xiaochun Nie, Thomas Pun, Roger Kumar, Hsi-Jung Wu
-
Publication number: 20080240600Abstract: Some embodiments of the invention provide a method of decoding an encoded video signal that includes a plurality of successive encoded images. The method initially receives an encoded image of the video signal. It then decodes the encoded image. The method next examines the decoded image to determine whether the decoded image satisfies at least one criterion for performing a post-filter operation on the decoded image. Only if the decoded image satisfies the criterion, the method performs the post-filter operation on the decoded image. The criterion in some embodiments is whether the amount of time for applying the post-filter operation exceeds the remaining amount of time that the method has to process the received image. Some embodiments of the invention provide a method of post-filtering a current image that was decoded from an encoded video signal, which includes a plurality of successive encoded images. The current image has several sub-sections. The method selects a sub-section of the current image.Type: ApplicationFiled: June 9, 2008Publication date: October 2, 2008Inventors: Thomas Pun, Maynard Handley, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20080232475Abstract: A method for encoding video with a two-dimensional (2D) transform separable to two one-dimensional (1D) transforms. The method receives an array of values for a sub-section of an image, performs a first 1D-transform of the array, transposes the resulting array, and performs a second 1D-transform of the array resulting from the transpose. The method, without performing another transpose, generates a data stream using a transposed scan order based on the values of the array resulting from the second transform. A method for decoding video encoded by a 2D transform, which separable to two 1D transforms. The method receives a data stream containing encoded values for an image, parses out the values into an array using a transposed scan order, performs a first 1D-inverse transform on the array, transposes the resulting array, and performs a second 1D-inverse transform of the array resulting from the transpose to produce a decoded output.Type: ApplicationFiled: April 28, 2008Publication date: September 25, 2008Inventors: Maynard Handley, Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20080232469Abstract: A rate controller for allocating a bit budget for video frames to be encoded is disclosed. The rate controller of the present invention considers many different factors when determining the frame bit budget including: desired video quality, target bit rate, frame type (intra-frame or inter-frame), frame duration, intra-frame frequency, frame complexity, intra-block frequency within an intra-frame, buffer overflow, buffer underflow, and the encoded video frame quality for a possible second pass.Type: ApplicationFiled: March 10, 2008Publication date: September 25, 2008Inventors: Xiaochun Nie, Thomas Pun, Roger Kumar, Hsi-Jung Wu
-
Patent number: 7418037Abstract: A rate controller for allocating a bit budget for video frames to be encoded is disclosed. The rate controller of the present invention considers many different factors when determining the frame bit budget including: desired video quality, target bit rate, frame type (intra-frame or inter-frame), frame duration, intra-frame frequency, frame complexity, intra-block frequency within an intra-frame, buffer overflow, buffer underflow, and the encoded video frame quality for a possible second pass.Type: GrantFiled: April 30, 2003Date of Patent: August 26, 2008Assignee: Apple Inc.Inventors: Xiaochun Nie, Thomas Pun, Roger Kumar, Hsi-Jung Wu
-
Patent number: 7412080Abstract: Some embodiments of the invention provide a method of performing motion estimation for an array of image values obtained from a current image. The method starts with an initial estimate of the motion of the current image. The method then determines whether the initial estimate of the motion of the current image array between the current image and a reference image is acceptable. If not, the method specifies a first value for a parameter for performing the motion estimation. Otherwise, the method specifies a second value for the parameter. The method then performs a motion estimation operation based on the specified value of the parameter. One example of a specified parameter is the size of the window that the motion estimation operation uses. Another example is the starting level for a hierarchical motion estimation operation that it performs. A hierarchical motion estimation operation searches the reference frame at several different levels of granularity.Type: GrantFiled: May 21, 2007Date of Patent: August 12, 2008Assignee: Apple IncInventors: Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7391812Abstract: Some embodiments of the invention provide a method of decoding an encoded video signal that includes a plurality of successive encoded images. The method initially receives an encoded image of the video signal. It then decodes the encoded image. The method next examines the decoded image to determine whether the decoded image satisfies at least one criterion for performing a post-filter operation on the decoded image. Only if the decoded image satisfies the criterion, the method performs the post-filter operation on the decoded image. The criterion in some embodiments is whether the amount of time for applying the post-filter operation exceeds the remaining amount of time that the method has to process the received image. Some embodiments of the invention provide a method of post-filtering a current image that was decoded from an encoded video signal, which includes a plurality of successive encoded images. The current image has several sub-sections. The method selects a sub-section of the current image.Type: GrantFiled: April 30, 2003Date of Patent: June 24, 2008Assignee: Apple Inc.Inventors: Thomas Pun, Maynard Handley, Roger Kumar, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20080147765Abstract: Some embodiments of the invention provide a method of performing a Discrete Cosine Transform (“DCT”) encoding or decoding coefficients of a data array by (1) multiplying the coefficients by a scalar value before the encoding or decoding, and then (2) dividing the encoded or decoded coefficients by the scalar value. When used in conjunction with fixed-point arithmetic, this method increases the precision of the encoded and decoded results. In addition, some embodiments provide a method of performing a two-dimensional (2D) Inverse Discrete Cosine Transform (“iDCT”). This method splits a pre-multiplication operation of the iDCT into two or more separate stages. When used in conjunction with fixed-point arithmetic, this splitting increases the precision of the decoded results of the iDCT.Type: ApplicationFiled: February 25, 2008Publication date: June 19, 2008Inventors: Roger Kumar, Maynard Handley, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7379956Abstract: Some embodiments of the invention provide a method of performing a Discrete Cosine Transform (“DCT”) encoding or decoding coefficients of a data array by (1) multiplying the coefficients by a scalar value before the encoding or decoding, and then (2) dividing the encoded or decoded coefficients by the scalar value. When used in conjunction with fixed-point arithmetic, this method increases the precision of the encoded and decoded results. In addition, some embodiments provide a method of performing a two-dimensional (2D) Inverse Discrete Cosine Transform (“iDCT”). This method splits a pre-multiplication operation of the iDCT into two or more separate stages. When used in conjunction with fixed-point arithmetic, this splitting increases the precision of the decoded results of the iDCT.Type: GrantFiled: April 30, 2003Date of Patent: May 27, 2008Assignee: Apple Inc.Inventors: Roger Kumar, Maynard Handley, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7376280Abstract: A method for encoding video with a two-dimensional (2D) transform separable to two one-dimensional (1D) transforms. The method receives an array of values for a sub-section of an image, performs a first 1D-transform of the array, transposes the resulting array, and performs a second 1D-transform of the array resulting from the transpose. The method, without performing another transpose, generates a data stream using a transposed scan order based on the values of the array resulting from the second transform. A method for decoding video encoded by a 2D transform, which separable to two 1D transforms. The method receives a data stream containing encoded values for an image, parses out the values into an array using a transposed scan order, performs a first 1D-inverse transform on the array, transposes the resulting array, and performs a second 1D-inverse transform of the array resulting from the transpose to produce a decoded output.Type: GrantFiled: April 30, 2003Date of Patent: May 20, 2008Assignee: Apple IncInventors: Maynard Handley, Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20070217514Abstract: Some embodiments of the invention provide a method of performing motion estimation for an array of image values obtained from a current image. The method starts with an initial estimate of the motion of the current image. The method then determines whether the initial estimate of the motion of the current image array between the current image and a reference image is acceptable. If not, the method specifies a first value for a parameter for performing the motion estimation. Otherwise, the method specifies a second value for the parameter. The method then performs a motion estimation operation based on the specified value of the parameter. One example of a specified parameter is the size of the window that the motion estimation operation uses. Another example is the starting level for a hierarchical motion estimation operation that it performs. A hierarchical motion estimation operation searches the reference frame at several different levels of granularity.Type: ApplicationFiled: May 21, 2007Publication date: September 20, 2007Inventors: Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Patent number: 7239721Abstract: Some embodiments of the invention provide a method of performing motion estimation for an array of image values obtained from a current image. The method starts with an initial estimate of the motion of the current image. The method then determines whether the initial estimate of the motion of the current image array between the current image and a reference image is acceptable. If not, the method specifies a first value for a parameter for performing the motion estimation. Otherwise, the method specifies a second value for the parameter. The method then performs a motion estimation operation based on the specified value of the parameter. One example of a specified parameter is the size of the window that the motion estimation operation uses. Another example is the starting level for a hierarchical motion estimation operation that it performs. A hierarchical motion estimation operation searches the reference frame at several different levels of granularity.Type: GrantFiled: April 30, 2003Date of Patent: July 3, 2007Assignee: Apple Inc.Inventors: Roger Kumar, Thomas Pun, Xiaochun Nie, Hsi-Jung Wu
-
Publication number: 20050286777Abstract: Some embodiments provide a method for encoding a first set of pixels in a first image by reference to a second image in a video sequence. In a first search window within a second image, the method searches to identify a first particular portion in the second image that best matches the first set of pixels in the first image. In the first search window within the second image, the method identifies a first location corresponding to the first particular portion. In a second search window within the second image, the method then searches to identify a second particular portion in the second image that best matches the first set of pixels in the first image, where the second search window is defined about the first location.Type: ApplicationFiled: April 28, 2005Publication date: December 29, 2005Inventors: Roger Kumar, Thomas Pun, Hsi Wu, Christian Duvivier