FIELD: physics.
SUBSTANCE: invention relates to image recognition. Technical result is achieved by the fact that the figures of people are detected on the current frame, the detected figure is highlighted by the corresponding bounding frame; performing bounding box positions prediction using Kalman filter; calculating the similarity of each predicted bounding box using a floristic commonality coefficient; generating a vector representation of the image enclosed in a bounding box for the figure; performing in the database matching of the generated vector representation of the figure by the nearest neighbour method; selecting a bounding box for the head within the detected bounding box for the figure; generating a vector representation of the image enclosed in a bounding box for the head; performing in the database matching of the generated head vector representation by the nearest neighbour method; and identifying a person by means of a weighted sum of results of matching by figure and by head.
EFFECT: high accuracy of identifying a person in a stream of video frames when there is overlapping of figures and heads.
3 cl
Title | Year | Author | Number |
---|---|---|---|
METHOD OF RECOGNIZING PEOPLE ON VIDEO IMAGES | 2023 |
|
RU2820757C1 |
METHOD AND SYSTEM FOR AUTOMATICALLY FINDING THREE-DIMENSIONAL IMAGES | 2005 |
|
RU2315352C2 |
SYSTEM AND METHOD FOR HUMAN BODY TEMPERATURE MONITORING BASED ON VIDEO DATA | 2020 |
|
RU2737138C1 |
METHOD AND SYSTEM FOR DETERMINING SYNTHETICALLY MODIFIED FACE IMAGES ON VIDEO | 2021 |
|
RU2768797C1 |
METHOD FOR TRACKING, DETECTION AND IDENTIFICATION OF OBJECTS OF INTEREST AND AUTONOMOUS DEVICE WITH PROTECTION FROM COPYING AND HACKING FOR THEIR IMPLEMENTATION | 2021 |
|
RU2789609C1 |
METHOD AND DEVICE FOR GENERATING VIDEO CLIP FROM TEXT DESCRIPTION AND SEQUENCE OF KEY POINTS SYNTHESIZED BY DIFFUSION MODEL | 2024 |
|
RU2823216C1 |
METHOD, TERMINAL AND SYSTEM FOR BIOMETRIC IDENTIFICATION | 2023 |
|
RU2815689C1 |
METHOD, TERMINAL AND SYSTEM FOR BIOMETRIC IDENTIFICATION | 2022 |
|
RU2798179C1 |
DETECTING OBJECTS FROM VISUAL SEARCH REQUESTS | 2017 |
|
RU2729956C2 |
METHOD OF RECOGNIZING NATURE OF TEXT CONTENT | 2023 |
|
RU2827987C1 |
Authors
Dates
2024-07-18—Published
2023-06-30—Filed