# US 2026/0300377 A1 · Server and method for labelling of a media file

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0300377 A1
- Document ID: US20260300377A1
- Status: Published application
- Kind: A1
- Application number: 19/577,618
- Filed: 2026-03-25
- Published: 2026-10-01
- Earliest claimed: 2025-03-26
- Assignees: Implid Ventures
- Applicants: Implid Ventures
- Inventors: Aymeric BRISSE; Stephane ROUCHERAY
- Web page: https://patentgrep.com/patent/US20260300377A1

## CPC

- G06F 16/45: Clustering; Classification
- G06F 16/41: Indexing; Data structures therefor; Storage structures
- G06F 40/205: Parsing
- G06F 40/295: Named entity recognition
- G06F 40/40: Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30)
- G10L 15/26: Speech to text systems (G10L15/08 takes precedence)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

Methods and servers for labelling of a media file are disclosed. The method includes accessing the media file, generating a first and a second entity tags for the media file respectively indicative of a presence of a first and a second entity detected in the media file between a first start timecode and a first end timecode and between a second start timecode and a second end timecode respectively. The method also includes a collision tag associating a collision between the first and second entity tags and comprising a presence of the first and second entities detected in the media file between a collision start timecode and a collision end timecode, the collision start time code being defined as the latest of the first and second start timecode, the collision end time code being defined as the earliest of the first and second end timecode.

## Claims

### Claim 1

A computer-implemented method for labelling of a media file, the method comprising:
accessing the media file ;
generating a first entity tag for the media file, the first entity tag comprising information about a first start timecode, a first end timecode and a presence of a first entity detected in the media file between the first start timecode and the first end timecode;
generating a second entity tag for the media file, the second entity tag comprising information about a second start timecode, a second end timecode and a presence of a second entity detected in the media file between the second start timecode and the second end timecode;
generating a collision tag associating a collision between the first and second entity tags, the collision tag being based on the first and second entity tags, the collision tag comprising a collision start time code, a collision end timecode and a presence of the first and second entities detected in the media file between the collision start timecode and the collision end timecode, the collision start time code being defined as the latest of the first and second start timecode, the collision end time code being defined as the earliest of the first and second end timecode; and
storing the collision tag in a database.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
receiving a retrieval instruction from a user, the retrieval instruction comprising a first target entity and a second target entity;
accessing the database; and
in response to determining that the first and the second target entities are associated with a given collision tag of the database, transmitting, to the user, a first indication associated with the given collision tag and a second indication associated with the given media file associated with the given collision tag.

### Claim 3 (depends on 2)

The method of claim 2, further comprising:
in response to determining that the first and the second target entities are associated with the given collision tag of the database, extracting a target extract from the media file between the collision start timecode and the collision end timecode; and
transmitting the target extract to the user.

### Claim 4 (depends on 2)

The method of claim 2, wherein:
generating the first entity tag comprises generating a textual description of the first entities in at least one natural language,
receiving a retrieval instruction comprises receiving a textual entry, and
determining that the first target entity is associated with the given collision tag of the database comprises determining that the textual entry matches at least partly the textual description of an entity associated with the given collision tag.

### Claim 5 (depends on 1)

The method of claim 1, wherein:
the media file comprises audio data, and
generating the first and second entity tags comprises employing a speech-to-text component on the audio data.

### Claim 6 (depends on 5)

The method of claim 5, wherein generating the first and second entity tags comprises:
parsing through a text file generated by the speech-to-text component; and
determining entities present in the media file based on the text file.

### Claim 7 (depends on 1)

The method of claim 1, wherein the first entity tag is associated with a type of entity tag selected in a group of entity tag types, said group comprising: audio entity tag type and video entity tag type.

### Claim 8 (depends on 1)

The method of claim 1, wherein:
the media file comprises video data, and
generating the first entity tag comprises:
employing a transcript component on the video data to generate a transcript text file representative of the video data under a textual form;
executing a Natural Language Processing machine learning algorithm to determine entities present in the media file.

### Claim 9 (depends on 8)

The method of claim 8, wherein the MLA is trained to perform Natural Language Processing tasks.

### Claim 10 (depends on 9)

The method of claim 9, wherein the MLA is trained to perform Named Entity Recognition tasks.

### Claim 11

A computer-implemented method for retrieving collisions in media files, the method comprising:
receiving a retrieval instruction from a user, the retrieval instruction comprising a first target entity and a second target entity;
accessing a database, the database comprising collision tags associating collisions, each of the collision tags comprising a collision start time code, a collision end timecode and a presence of entities detected in the media file between the collision start timecode and the collision end timecode; and
in response to determining that the first and the second target entities are associated with a given collision tag of the database, transmitting, to the user, a first indication associated with the given collision tag and a second indication associated with the given media file associated with the given collision tag.

### Claim 12 (depends on 11)

The method of claim 11, wherein transmitting, to the user, the first indication associated with the given collision tag and the second indication associated with the given media file associated with the given collision tag comprises displaying, on a graphical user interface (GUI), the first indication and the second indication along with a visual representation indicative of the given media file.

### Claim 13 (depends on 11)

The method of claim 11, wherein the collision tag has been generated by executing, for each of the media files:
generating a first entity tag, the first entity tag comprising information about a first start timecode, a second end timecode and a presence of a first entity detected between the first start timecode and the second end timecode;
generating a second entity tag, the second entity tag comprising information about, a second start timecode, a second end timecode and a presence of a second entity detected between the second start timecode and the second end timecode; and
generating the collision tag associating a collision between the first and second entity tags, the collision tag comprising a collision start time code, a collision end timecode and a presence of the first and second entities detected between the collision start timecode and the collision end timecode, the collision start time code being defined as the latest of the first and second start timecode, the collision end time code being defined as the earliest of the first and second end timecode.

### Claim 14 (depends on 1)

A system for concomitant labelling of a media file, the system comprising a controller and a memory storing a plurality of executable instructions which, when executed by the controller, cause the system to perform the computer-implemented method of claim 1.

### Claim 15 (depends on 1)

A non-transitory computer-readable medium comprising computer-readable instructions that, upon being executed by a system, cause the system to perform the computer-implemented method of claim 1.
