Topic outline

  • Image Recognition model training

    Visual Testing Program, Acton, MA. 

    • 2 QA entry level testers and 1 team leader 
    • 4 weeks about 10 hours per week  
    • Visual Verification of automated Image Recognition results 
    • Detailed reporting and analysis
    • Training and testing image recognition model

      Training model is a method to quickly evaluate and adjust the performance of the algorithm. The training dataset is used to train the model. The test dataset is considered as new data where the output values are withheld from the algorithm.

      WeCanTest verifies results from the trained model on the inputs from the test dataset. Then these results are compared to the withheld output of the test set. Comparing the WeCanTest test results and withheld outputs on the test dataset provides allows to correct weights and measures of the model.

      WeCanTest participants work on training Siena Analytics deep learning algorithms verifying results of the image recognition test datasets. It helps Siena Analytics efficiently analyse relevant data and images from logistics automation equipment including:
      • Machine data from all sensors and camera tunnel equipment
      • Package data and images
      • Label images and details
      and presents it in actionable dashboards – where and when it’s needed.