cricket-tannmay-headless 0.1.1

A computer vision package to recognize and assess different cricket shots and bowling videos.

Cricket

This is a package which can be used for assessing videos of bowling and batting.It is different from cricket as this can work in situations where opencv-python is throwing this type of error:

ImportError: libGL.so.1: cannot open shared object file: No such file or directory

Note: Please add a packages.txt file inside your project having the following data:

libgl1-mesa-glx
libglib2.0-0
libxrender1

Installation

Option 1

Use the command:

pip install cricket-tannmay

Option 2

Go to this link. Then download the source code and requirements.txt and run this command:

pip install -r requirements.txt

How To Use

Note: The name to use while installing is cricket-tannmay but while using it in python we have to import cricket

Batting

Specific Shot

Option 1

This option will give a single string in return telling It will also if it was a:

• Good shot
• Bad shot
• Not the shot the function was called for

This is the implementation for the compatible shots:

• Straight Drive:

from cricket.batting import straight_drive
decision = straight_drive.classify_entire_video("video_path")

• Cover drive:

from cricket.batting import cover_drive
decision = cover_drive.classify_entire_video("video_path")

• Pull shot:

from cricket.batting import pull_shot
decision = pull_shot.classify_entire_video("video_path")

• Sweep:

from cricket.batting import sweep
decision = sweep.classify_entire_video("video_path")

• Reverse sweep:

from cricket.batting import reverse_sweep
decision = reverse_sweep.classify_entire_video("video_path")

• Hook shot:

from cricket.batting import hook_shot
decision = hook_shot.classify_entire_video("video_path")

• Defense:

from cricket.batting import defense
decision = defense.classify_entire_video("video_path")

Option 2

This will return a dictionary which tells about what it thinks at that frame, whether it is a:

• Good Shot
• Bad Shot
• Not the shot it was called for

This is the implementation for the compatible shots:

• Straight Drive:

from cricket.batting import straight_drive
decision = straight_drive.analyze_video("video_path")

• Cover drive:

from cricket.batting import cover_drive
decision = cover_drive.analyze_video("video_path")

• Pull shot:

from cricket.batting import pull_shot
decision = pull_shot.analyze_video("video_path")

• Sweep:

from cricket.batting import sweep
decision = sweep.analyze_video("video_path")

• Reverse sweep:

from cricket.batting import reverse_sweep
decision = reverse_sweep.analyze_video("video_path")

• Hook shot:

from cricket.batting import hook_shot
decision = hook_shot.analyze_video("video_path")

• Defense:

from cricket.batting import defense
decision = defense.analyze_video("video_path")

Overall

from cricket.batting import overall_shot_detection
classifier = overall_shot_detection.CricketShotClassifier()
classifier.classify_entire_video("video_path")

This classifier class is compatible with the following shots:

• Straight Drive
• Cover Drive
• Sweep
• Defence
• Reverse Sweep
• Pull Shot
• Hook Shot

This will also return if it was a:

• Good Shot
• Bad Shot

Balling

Specific Type

from cricket.balling import fast
import cricket.utils as utils

ball_colour_range = utils.color_ranges["white"]
decision = fast.classify_entire_video("video_path", ball_colour_range)

This method is compatible with:

• Fast ball
• Slow ball
• Yorker
• Spin

This was an implementation of a fast ball. It will also tell if it was a:

• Good Delivery
• Bad Delivery

Overall

from cricket.balling.overall_bowl_detection import CricketBallAnalyzer
import cricket.utils as utils

ball_colour_range = utils.color_ranges["white"]
analyser = CricketBallAnalyzer("video_path", ball_colour_range[0], ball_colour_range[1], 22)
decision = analyser.analyze_video()

This classifier class is compatible with:

• Fast Ball
• Slow ball
• Yorker
• Spin

Combined

from cricket.combined import classify_overall
import cricket.utils as utils

ball_colour = utils.color_ranges["white"]
decision = classify_overall("video_path", ball_colour, 22)

This will run both the classes given above and will return a dictionary. First key will be of bowling and second key will be of batting

Note: For ease of use you can import like this and it would import all things for implementation from both the subpackages

from cricket import *