A trained deep learning model able to detect defects in QA images of a production line, packaged in a ready-to-deploy Docker image together with a web front-end where pictures can be uploaded for classification.
The images show a surface with potential defects. Labeling data is provided as JSON in labelME format, containing bounding boxes for two classes of defects. A single image shows tiled sub-images and may contain multiple defects. The dataset contains approximately 1500 images.
The model must reach >95% accuracy, as measured on a test set that will not be provided. Low false negative (error not detected) is more important than low false positive rate.
The Docker image must contain a web frontend (simple form) where images may be uploaded for detection. Multiple images, as well as ZIP files containing multiple images, must be uploadable. The resulting output images must be the input image with sub-tiles containing defects highlighted and showing the detection confidence. A checkbox must be provided to allow the user to decide if the resulting images will be displayed directly in the browser or downloaded as a single [login to view URL] run a Data Science Academy and require a Data Science practitioner to create a series of Data Science videos for us.
We have all the content created already (in Jupyter Notebooks) and require the resource to create videos from these.
The resource needs to:
- Be able to explain technical content to beginners to the subject
- Have perfect, clear English
- Be able to produce videos (have the necessary equipment to do this)
- Have a fun, animated personality to bring the videos to life
- Be a solid Data Science practitioner and understand all the foundations (eg. Python, SQL, PowerBI and the Supervised and Unsupervised algorithms)
The ideal candidate is young, smart and has studied Data Science recently - so the concepts are fresh in their minds
9 фрилансеров(-а) в среднем готовы выполнить эту работу за $19/час
Dear Sir, I have more than 3 years experience in Data science. I am very familiar with machine learning and deep learning. I will love to be considered for. this role. Regards.