TDDS is an AI-powered system which takes an X-ray video of a sample tire and outputs a float number between 0 and 1.0 indicating the defect precentage of the sample tire.
The CNN used to obtain the defect percentage cannot process the raw input video because of two reasons :
- The size of the video is too much and hence can not be processed in real-time. In fact it could take months to even train such a model.
- There is large amount of repeated and redundant pixels in tandem frames especially if the input video is captured with a high FPS.
Thus, it is neither logical nor optimal to use the video as input.
Input :
output :
