Methods and systems for extracting sport-related information from digital video frames

1. A method for automatically extracting, from a digital video frame, scoreboard information including a first team name, a second team name, a first score, and a second score, the method comprising:

detecting by a computing system, within the digital video frame, a plurality of frame regions based on each detected frame region depicting text, wherein the computing system comprises a processing unit;

selecting by the computing system, from the detected frame regions, a set of frame regions based on the frame regions of the selected set cooperatively having a geometric arrangement that corresponds with a candidate geometric arrangement of the scoreboard information, wherein selecting the set comprises selecting frame regions of the set based on the selected frame regions of the set having particular geometric characteristics, and wherein selecting frame regions of the set based on the selected frame regions having particular geometric characteristics comprises selecting frame regions of the set based on first geometric characteristics of one of the selected frame regions substantially matching second geometric characteristics of another one of the selected frame regions;

recognizing, by the computing system, characters respectively within each of the frame regions of the selected set of frame regions;

based at least on the recognized characters in the frame regions of the selected set, detecting by the computing system the scoreboard information, including (i) detecting the first team name based on the recognized characters in a first frame region of the selected set, (ii) detecting the second team name based on the recognized characters in a second frame region of the selected set, (iii) detecting the first score based on the recognized characters in a third frame region of the selected set, and (iv) detecting the second score based on the recognized characters in a fourth frame region of the selected set; and

recording by the computing system the detected scoreboard information.

2. The method of claim 1 , further comprising:

receiving a real-time broadcast feed including the digital video frame,

wherein automatically extracting the scoreboard information comprises automatically extracting the scoreboard information while receiving the real-time broadcast feed.

3. The method of claim 1 , further comprising:

determining, by the computing system, a scoreboard region of the digital video frame,

wherein detecting the plurality of frame regions comprises detecting, within the determined scoreboard region of the digital video frame, the plurality of frame regions based on each detected frame region depicting text.

4. The method of claim 1 , wherein selecting the set is further based on the selected set including a predetermined number of frame regions that cooperatively have the geometric arrangement.

5. The method of claim 4 , wherein the predetermined number of frame regions is four frame regions.

6. The method of claim 1 , wherein the digital video frame depicts a sport event, the method further comprising:

determining, by the computing system, the sport event depicted by the digital video frame; and

based on the determined sport event, determining, by the computing system, a number of frame regions to be included in the set of frame regions,

wherein selecting the set is further based on the selected set including the determined number of frame regions.

7. The method of claim 1 , wherein the candidate geometric arrangement is a rectangular arrangement.

8. The method of claim 1 ,

wherein the first geometric characteristics comprise one or more of (i) a first height, (ii) a first width, (iii) a first size, and (iv) a first shape, and

wherein the second geometric characteristics comprise one or more of (i) a second height, (ii) a second width, (iii) a second size, and (iv) a second shape.

9. The method of claim 1 ,

wherein the computing system has access to data indicating a plurality of candidate locations associated with the candidate geometric arrangement, including (i) a candidate location of the first team name in the candidate geometric arrangement, (ii) a candidate location of the second team name in the candidate geometric arrangement, (iii) a candidate location of the first score in the candidate geometric arrangement, and (iv) a candidate location of the second score in the candidate geometric arrangement, and

wherein detecting the scoreboard information is further based on the data indicating the plurality of candidate locations associated with the candidate geometric arrangement.

10. The method of claim 9 , wherein each frame region of the set has a respective location, in the set, relative to other frame regions of the set, and wherein detecting the scoreboard information based on the data indicating the plurality of candidate locations associated with the candidate geometric arrangement comprises one or more of:

(i) determining that the respective location of the first frame region in the set corresponds to the candidate location of the first team name in the candidate geometric arrangement, and responsively detecting the first team name based on the recognized characters in the first frame region,

(ii) determining that the respective location of the second frame region in the set corresponds to the candidate location of the second team name in the candidate geometric arrangement, and responsively detecting the second team name based on the recognized characters in the second frame region,

(iii) determining that the respective location of the third frame region in the set corresponds to the candidate location of the first score in the candidate geometric arrangement, and responsively detecting the first score based on the recognized characters in the third frame region, and

(iv) determining that the respective location of the fourth frame region in the set corresponds to the candidate location of the second score in the candidate geometric arrangement, and responsively detecting the second score based on the recognized characters in the fourth frame region.

11. The method of claim 1 , wherein detecting the scoreboard information based on the recognized characters in the frame regions of the selected set comprises one or more of:

(i) determining that the first frame region includes just letters as the recognized characters, and responsively detecting the first team name based on the recognized characters in the first frame region,

(ii) determining that the second frame region includes just letters as the recognized characters, and responsively detecting the second team name based on the recognized characters in the second frame region,

(iii) determining that the third frame region includes just one or more digits as the recognized characters, and responsively detecting the first score based on the recognized characters in the third frame region, and

(iv) determining that the fourth frame region includes just one or more digits as the recognized characters, and responsively detecting the second score based on the recognized characters in the fourth frame region.

12. The method of claim 1 , further comprising:

making a determination that (i) two frame regions of the set each respectively include just letters as the recognized characters and (ii) at least two other frame regions of the set each respectively include just one or more digits as the recognized characters,

wherein the detecting of the scoreboard information being based on the recognized characters in the frame regions of the selected set is responsive to making the determination.

13. The method of claim 1 , wherein the recognized characters are recognized first characters, wherein the digital video frame depicts a sport event, and wherein the scoreboard information also includes a game time associated with the sport event, the method further comprising:

recognizing, by the computing system, second characters within a particular frame region of the plurality of frame regions;

making a determination, by the computing system, that the second characters recognized within the particular frame region are arranged according to a predetermined pattern; and

in response to making the determination, detecting, by the computing system, the game time based on (i) the second characters recognized within the particular frame region and (ii) the predetermined pattern.

14. The method of claim 13 , wherein the predetermined pattern is a regular expression predefining a sequence of candidate characters for representing the game time.

15. A computing system comprising:

a processing unit;

non-transitory data storage; and

program instructions stored in the non-transitory data storage and executable by the processing unit to carry out operations for automatically extracting, from a digital video frame, scoreboard information including a first team name, a second team name, a first score, and a second score, the operations comprising:

detecting, within the digital video frame, a plurality of frame regions based on each detected frame region depicting text,

selecting, from the detected frame regions, a set of frame regions based on the frame regions of the selected set cooperatively having a geometric arrangement that corresponds with a candidate geometric arrangement of the scoreboard information, wherein selecting the set comprises selecting frame regions of the set based on the selected frame regions of the set having particular geometric characteristics, and wherein selecting frame regions of the set based on the selected frame regions having particular geometric characteristics comprises selecting frame regions of the set based on first geometric characteristics of one of the selected frame regions substantially matching second geometric characteristics of another one of the selected frame regions,

recognizing characters respectively within each of the frame regions of the selected set of frame regions,

based at least on the recognized characters in the frame regions of the selected set, detecting the scoreboard information, including (i) detecting the first team name based on the recognized characters in a first frame region of the selected set, (ii) detecting the second team name based on the recognized characters in a second frame region of the selected set, (iii) detecting the first score based on the recognized characters in a third frame region of the selected set, and (iv) detecting the second score based on the recognized characters in a fourth frame region of the selected set, and

recording the detected scoreboard information.

16. The computing system of claim 15 , wherein the candidate geometric arrangement is a rectangular arrangement.

17. The computing system of claim 16 , wherein selecting the set is further based on the selected set including four frame regions that cooperatively have the rectangular arrangement.

18. The computing system of claim 15 ,

wherein the first geometric characteristics comprise one or more of (i) a first height, (ii) a first width, (iii) a first size, and (iv) a first shape, and

wherein the second geometric characteristics comprise one or more of (i) a second height, (ii) a second width, (iii) a second size, and (iv) a second shape.

19. A non-transitory computer readable medium having stored thereon instructions executable by a processing unit to cause a computing system to perform operations for automatically extracting, from a digital video frame, scoreboard information including a first team name, a second team name, a first score, and a second score, the operations comprising:

detecting, within the digital video frame, a plurality of frame regions based on each detected frame region depicting text;

selecting, from the detected frame regions, a set of frame regions based on the frame regions of the selected set cooperatively having a geometric arrangement that corresponds with a candidate geometric arrangement of the scoreboard information, wherein selecting the set comprises selecting frame regions of the set based on the selected frame regions of the set having particular geometric characteristics, and wherein selecting frame regions of the set based on the selected frame regions having particular geometric characteristics comprises selecting frame regions of the set based on first geometric characteristics of one of the selected frame regions substantially matching second geometric characteristics of another one of the selected frame regions;

recognizing characters respectively within each of the frame regions of the selected set of frame regions;

based at least on the recognized characters in the frame regions of the selected set, detecting the scoreboard information, including (i) detecting the first team name based on the recognized characters in a first frame region of the selected set, (ii) detecting the second team name based on the recognized characters in a second frame region of the selected set, (iii) detecting the first score based on the recognized characters in a third frame region of the selected set, and (iv) detecting the second score based on the recognized characters in a fourth frame region of the selected set; and

recording the detected scoreboard information.

20. The non-transitory computer readable medium of claim 19 ,

wherein the first geometric characteristics comprise one or more of (i) a first height, (ii) a first width, (iii) a first size, and (iv) a first shape, and

wherein the second geometric characteristics comprise one or more of (i) a second height, (ii) a second width, (iii) a second size, and (iv) a second shape.