Image Object Recognition based on Deep Neural Network using Keras

Job ID: 33593006

Budget: $170 – $240 USD

The Goal:

Detect and Count number of teeth on the excavator bucket - see image attached.

The Task:

1. Take video dataset (6 videos) provided by me and label it using feely available tools (LabelMe, LabelImg).
2. Build a deep neural network (semantic segmentation or object detection) to recognize excavator bucket teeth.
3. Train the network.
4. Test and show detection results as on the image attached.

Deliverables:

1. Conda/Python environment specification.
2. Python script that loads labeled data, defines and trains the network
3. Labeled dataset and instructions how to label
4. Python scripts that runs and detects teeth on the new video file.