Granted Patent B2

US 12,573,093 · App. 18/170,060 · Granted Mar 10, 2026

Inventors: Chuan Li (Troy, MI); Andrew W. Averhart (Redford, MI); Fan Bai (Ann Arbor, MI)

Assignee: GM GLOBAL TECHNOLOGY OPERATIONS LLC

G06T9/00G06T5/70G06V10/764G06V2201/07

Abstract

Claims (58)

1 . A system for object blurring in compressed videos, the system comprising:

one or more sensors;

one or more human machine interfaces (HMIs);

one or more control modules, each of the one or more control modules having a processor, a memory, and one or more input/output (I/O) ports, the one or more control modules in communication with the one or more sensors and in communication with the one or more HMIs via the I/O ports, the memory storing programmatic control logic, the processor executing the programmatic control logic, the programmatic control logic comprising:

a first control logic that receives, via the I/O ports, raw input image data from the one or more sensors;

a second control logic that encodes the raw input image data into compressed image data and stores the compressed image data in memory;

a third control logic that parses and partially decodes the compressed image data to access raw independent video frames, including:

control logic that identifies independent video frames within the compressed image data;

control logic that decodes the independent video frames within the compressed video data and outputs the raw independent video frames;

control logic for one or more object detection algorithms, the one or more object detection algorithms executed on the raw independent video frames;

control logic that classifies objects detected within the raw independent video frames; and

upon determining that an object within the raw independent video frames is within a protected class, executes an obfuscation control logic on the object in the protected class; and

upon determining that an object in the raw independent video frames is not within a protected class, does not execute the obfuscation control logic with respect to the object not in a protected class;

a fourth control logic that performs one or more post-processing algorithms on the raw independent video frames to generate post-processed independent video frames;

a fifth control logic that encodes the post-processed independent video frames, and replaces previously encoded raw independent video frames in the compressed image data with the encoded post-processed independent video frames; and

a sixth control logic that transmits, via the I/O ports, an output to the HMI, wherein the output includes the encoded post-processed independent video frames, and a plurality of dependent video frames, wherein the plurality of dependent video frames depend from the encoded post-processed independent video frames.

2 . The system of claim 1 , wherein the one or more sensors comprise:

one or more cameras; and

wherein the one or more HMIs further comprise: a screen visible by a human.

3 . The system of claim 1 , wherein the obfuscation control logic alters pixel values of the object in the protected class, thereby at least partially obfuscating the object in the protected class.

4 . The system of claim 1 , wherein the fifth control logic encodes the post-processed independent video frames including obfuscated objects; and

overwrites previous encoded raw independent video frames in the compressed image data with the encoded post-processed independent video frames.

5 . The system of claim 4 , wherein the sixth control logic that transmits an output to the HMI, wherein the output is in a group of pictures (GOP) concept that includes the encoded post-processed independent video frames and the plurality of dependent video frames.

6 . The system of claim 5 , wherein dependent video frames include time and spatial reference information that associates each of the plurality of dependent video frames with an independent video frame, and wherein upon decoding the dependent video frames, obfuscated objects in the independent video frames are also associated with corresponding dependent video frames.

7 . A method for object blurring in compressed videos, the method comprising:

capturing raw input image data with one or more cameras;

utilizing one or more control modules, each of the one or more control modules having a processor, a memory, and one or more input/output (I/O) ports, the one or more control modules in communication with the one or more cameras and in communication with one or more human-machine interfaces (HMIs) via the I/O ports, the memory storing programmatic control logic, the processor executing the programmatic control logic for:

encoding the raw input image data into compressed image data and stores the compressed image data in memory;

parsing and partially decoding the compressed image data to access raw independent video frames, including:

identifying independent video frames within the compressed image data;

decoding the independent video frames within the compressed video data and outputs the raw independent video frames;

executing one or more object detection algorithms, the one or more object detection algorithms executed on the raw independent video frames;

classifying objects detected within the raw independent video frames;

upon determining that an object within the raw independent video frames is within a protected class, executing an obfuscation control logic on the object in the protected class; and

upon determining that objects in the raw independent video frames are not within a protected class, not executing the obfuscation control logic with respect to the object not in a protected class;

performing one or more post-processing algorithms on the raw independent video frames to generate post-processed independent video frames;

encoding the post-processed independent video frames, and replaces previously encoded raw independent video frames in the compressed image data with the encoded post-processed independent video frames; and

transmitting, via the I/O ports, an output to the one or more HMIs, wherein the output includes the encoded post-processed independent video frames, and a plurality of dependent video frames, wherein the plurality of dependent video frames depend from the encoded post-processed independent video frames.

8 . The method of claim 7 , further comprising:

altering pixel values of the object in the protected class, thereby at least partially obfuscating the object in the protected class.

9 . The method of claim 7 , further comprising:

encoding the post-processed independent video frames including obfuscated objects; and

overwriting previous encoded raw independent video frames in the compressed image data with the encoded post-processed independent video frames.

10 . The method of claim 9 , wherein transmitting an output further comprises:

transmitting a group of picture (GOP) output that includes the encoded post-processed independent video frames and a plurality of dependent video frames.

11 . The method of claim 9 , wherein upon decoding the dependent video frames, obfuscated objects in the independent video frames are also associated with corresponding dependent video frames utilizing time and spatial reference information that associates each of the plurality of dependent video frames with an independent video frame.

12 . A system for object blurring in compressed videos, the system comprising:

one or more cameras;

one or more human machine interfaces (HMIs), each of the one or more HMIs having a screen visible by humans;

one or more control modules, each of the one or more control modules having a processor, a memory, and one or more input/output (I/O) ports, the one or more control modules in communication with the one or more cameras and in communication with the one or more HMIs via the I/O ports, the memory storing programmatic control logic, the processor executing the programmatic control logic, the programmatic control logic comprising:

a first control logic that receives, via the I/O ports, raw input image data from the one or more sensors;

a second control logic that encodes the raw input image data into compressed image data and stores the compressed image data in memory;

a third control logic that parses and partially decodes the compressed image data to access raw independent video frames by identifying independent video frames within the compressed image data and decoding the independent video frames within the compressed video data, wherein the third control logic further includes control logic to output the raw independent video frames and to execute one or more object detection algorithms on the raw independent video frames; and control logic that classifies objects detected within the raw independent video frames; and upon determining that an object within the raw independent video frames is within a protected class, executes an obfuscation control logic on the object in the protected class; and upon determining that an object in the raw independent video frames is not within a protected class, does not execute the obfuscation control logic with respect to the object not in a protected class;

a fourth control logic that performs one or more post-processing algorithms on the raw independent video frames to generate post-processed independent video frames;

a fifth control logic that encodes the post-processed independent video frames, and replaces and overwrites previously encoded raw independent video frames in the compressed image data with the encoded post-processed independent video frames; and

a sixth control logic that transmits, via the I/O ports, an output to the HMI, wherein the output includes the encoded post-processed independent video frames, and a plurality of dependent video frames, wherein the plurality of dependent video frames depend from the encoded post-processed independent video frames.

13 . The system of claim 12 , wherein the obfuscation control logic alters pixel values of the object in the protected class, thereby at least partially obfuscating the object in the protected class.

14 . The system of claim 12 , wherein the sixth control logic that transmits an output to the HMI, wherein the output is in a group of picture (GOP) concept that includes the encoded post-processed independent video frames and a plurality of dependent video frames, wherein the dependent video frames include time and spatial reference information that associates each of the plurality of dependent video frames with an independent video frame, and wherein upon decoding the dependent video frames, obfuscated objects in the independent video frames are also associated with corresponding dependent video frames.

Assignments (1)

Continuity (1)

References Cited (43)