CLASS ACCURACY RESULTS:
- Accuracy Metric #1:
This metric measures how many truss strokes your recognizer correctly found for each truss sketch. In other words:
- (# of model truss strokes your recognizer found) / (# of model truss strokes)
- Accuracy Metric #2:
This metric measures how many of the truss strokes correctly found are actually in the model truss strokes. In other words:
- (# of truss strokes your recognizer found that are also in model truss strokes) / (# of truss strokes your recognizer found)
- Overall Accuracies:
This measurement lists the averages of both the first and second accuracy metrics. In other words:
- (sum of accuracies) / (number of data files)
DEADLINE:
Code: Thursday, Sep. 30th @ 11:59 PM
Report: Tuesday, Oct. 5th @ start of class (paper form)
DESCRIPTION:
This project will involve recognizing trusses. You will first provide sketching data of the trusses and label them. Afterwards, you will build a recognizer using a library of corner-finding algorithms for recognizing trusses.
REQUIREMENTS:
This programming project consists of at least three steps. The requirements will be updated throughout the week.
EXTRA INFORMATION:
This section will give you more information to begin the assignment, especially if you are new to the tools used in sketch recognition.
