Granted Patent B2

US 12,555,612 · App. 18/830,767 · Granted Feb 17, 2026

Inventors: Chul Soon Park (Hanam-si, KR); Byeong Nam Ahn (Yongin-si, KR)

Assignee: WISEJET INC.

G11B27/36G11B27/34H04N21/44008H04N21/44209

Abstract

Claims (44)

1 . A method for transmitting and receiving video based on wireless transmission in a reception device, the method comprising:

receiving video frames that are wirelessly transmitted from a transmission device using a predetermined frequency band;

detecting an error in at least one video frame of the received video frames; and

correcting the error in the at least one video frame with the detected error (hereafter, error video frame) by inputting the error video frame, together with a previous video frame, into a pre-trained artificial intelligence (AI) model,

wherein the detecting an error in the at least one video frame comprises:

monitoring a bit error rate of network environment of the wireless transmission; and

determining that a video frame includes an error when the bit error rate exceeds a preset threshold value, and

wherein the correcting of the error in the error video frame comprises:

requesting the transmission device to temporarily store a plurality of video frames from a point in time at which the error was detected;

requesting the transmission device to transmit the plurality of temporarily stored video frames and a current video frame at a point in time at which the bit error rate becomes less than or equal to the preset threshold value; and

sequentially playing the plurality of temporarily stored video frames and the current video frame.

2 . The method of claim 1 , wherein the correcting of the error of the error video frame comprises:

generating an error generation region in the error video frame as lost information;

generating correction information from a region corresponding to the lost information in the previous video frame; and

correcting the region corresponding to the lost information in the error video frame on the basis of the correction information.

3 . The method of claim 2 , wherein the generating of correction information from a region corresponding to the lost information in the previous video frame comprises:

identifying an object region and a background region that correspond to the lost information of a plurality of previous video frames; and

generating position information of each of the object region and the background region and the amount of variation of the position information as the correction information.

4 . The method of claim 2 , wherein the generating of correction information from a region corresponding to the lost information in the previous video frame comprises: when receiving a plurality of error video frames,

identifying an object region and a background region that correspond to the lost information in each of a previous video frame received before the plurality of error video frames and a current video frame received after the plurality of error video frames; and

generating each position information of an object region and a background region in the previous video frame and the current video frame, and the amount of variation of the position information as the correction information.

5 . The method of claim 2 , wherein the correcting of the error of the error video frame comprises: when correcting a second error video frame after the correction of a first error video frame is completed,

calculating similarity of lost information between the first error video frame and the second error video frame; and

applying correction information of the first error video frame to the second error video frame when the similarity of lost information is greater than or equal to a preset threshold value.

6 . The method of claim 2 , wherein the correcting of the error of the error video frame comprises: when correcting a second error video frame after the correction of a first error video frame is completed,

calculating similarity of lost information between the first error video frame and the second error video frame;

discriminating a coincident part and a non-coincident part of error generation regions of the first error video frame and the second error video frame when the similarity of lost information is less than the preset threshold value;

applying first correction information of the first error video frame to the second error video frame for the coincident part; and

generating second correction information for the second error video frame for the non-coincident part.

7 . The method of claim 1 , wherein the correcting of the error of the error video frame is to apply the currently trained AI model to a next unit period when a correction ratio calculated between the error video frame and the corrected video frame is greater than or equal to a preset minimum correction ratio.

8 . The method of claim 1 , further comprising:

retraining the AI model when a correction ratio calculated between the error video frame and the corrected video frame is less than a preset minimum correction ratio; and

applying the retrained AI model to a next unit period.

9 . A reception device, comprising:

a communication module receiving video frames that are wirelessly transmitted from a transmission device using a predetermined frequency band;

a memory storing a program for detecting and correcting an error of at least one video frame of the received video frames; and

a processor correcting the error of the video frame with the detected error (hereafter, error video frame) by inputting the error video frame into a pre-trained AI model together with a previous video frame as the program stored in the memory is executed,

wherein the processor monitors a bit error rate for the wirelessly transmitted network environment and detects the bit error rate as an error of the video frame when the bit error rate is detected as exceeding a preset threshold value,

wherein the processor requests the transmission device to temporarily store a plurality of video frames from a point in time at which the error was detected, requests the transmission device to transmit the plurality of temporarily stored video frames and a current video frame at a point in time at which the bit error rate becomes less than or equal to the preset threshold value, and then sequentially plays the plurality of temporarily stored video frames and the current video frame.

10 . The reception device of claim 9 , wherein the processor generates an error generation region in the error video frame as lost information, generates correction information from a region corresponding to the lost information in the previous video frame, and corrects the region corresponding to the lost information in the error video frame on the basis of the correction information.

11 . The reception device of claim 10 , wherein the processor identifies an object region and a background region that correspond to the lost information of a plurality of previous video frames, and generates position information of each of the object region and the background region and the amount of variation of the position information as the correction information.

12 . The reception device of claim 10 , wherein when receiving a plurality of error video frames, the processor identifies an object region and a background region that correspond to the lost information in each of a previous video frame received before the plurality of error video frames and a current video frame received after the plurality of error video frames, and generates each position information of an object region and a background region in the previous video frame and the current video frame, and the amount of variation of the position information as the correction information.

13 . The reception device of claim 10 , wherein when correcting a second error video frame after the correction of a first error video frame is completed, the processor calculates similarity of lost information between the first error video frame and the second error video frame, and applies correction information of the first error video frame to the second error video frame when the similarity of lost information is greater than or equal to a preset threshold value.

14 . The reception device of claim 9 , wherein when correcting a second error video frame after the correction of a first error video frame is completed, the processor calculates similarity of lost information between the first error video frame and the second error video frame, discriminates a coincident part and a non-coincident part of error generation regions of the first error video frame and the second error video frame when the similarity of lost information is less than the preset threshold value, applies first correction information of the first error video frame to the second error video frame for the coincident part, and generates second correction information for the second error video frame for the non-coincident part.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (19)