Bicycle crowd evaluation

Bicycle Crowd evaluation Project. This is for the evaluation of the new crowd in the automotive dataset given at (https://www.a2d2.audi).

For the tasks there are 4 main task + presentation.

Now to make it easy have created 4 folders for each of the task.

Now the task1 is the folder for Task-1. This task has 4 sub parts.

• Task1a:

  • list the annotators for the bicycle dataset. Just run the file with command python Task1a.py to see the output.

• Task1b:

  • This is to get the annotations durations that is min, max duration times. Just run the file with command python Task1b.py to see the output.

• Task1c:

  • This is to get the difference between the annotations result. Just run the file with command python Task1c.py to see the output.

• Task1d:

  • This is to give the annotators that disaggree the most. Just run the file with command python Task1d.py to see the output

Now the task2 is the folder for Task-2.

• Task2:
  • This task is to analyze the annotations with label corrupt and can't solve. Just run the file with command python task2.py to see the output

Now the task3 is the folder for Task-3.

• Task3:
  • This task is to analyze whether the reference set is balanced or not. Just run the file with command python task3.py to see the output.

Now the task4 is the folder for Task-4.

• Task4:
  • This task is to analyze good and bad annotators for the given dataset. All plots are given. Just run the file with command python task4.py to see the output.

Now task 5 is presentation. This is there in the presentation folder. This gives the summary of the findings.