Drill bit identification task

Problem Statement

• From the image of a drill bit, identify the drill size. This task has been solved as a supervised learning - classification problem wherein we train the model with different frames and provide it the expected drill sizes. During testing, we provide the frame to the model and predict which drill bit size it belongs to.

Windows executable - How to run model inference with inference.exe

• Please place all unseen images in the inference_images directory and execute the inference.exe file

Python - How to run model inference with inference.py

• Within a conda virtual environment please install

  • pytorch 1.13
  • python 3.9
  • opencv-python

    conda create -n env_setup python=3.9
    pip install opencv-python
    conda install pytorch torchvision torchaudio -c pytorch (For MAC only)
    

• Once done, please place all unseen images in the inference_images directory and run

  python inference.py
  

• Since the images would be placed in the inference_images folder, the program automatically picks up the images and generates predictions. These predictions will be written to a text file - "output.txt"

• A sample output.txt (generated on the images present inside inference_images) has been provided to see what the final output would look like.

Note - Python only.

One can also provide another location to the unseen images by executing

python inference.py --inferenceDataset './inference_images/'