Tell us about your project!
Agriculture
Highly-specialized team of biologists helped to classify plant diseases for timely crop monitoring.
Training data for AI and machine learning in agriculture
About case:
About case:
Industry and use case
Agriculture, plant-disease detection
DATA
50,000 plant photos
Project duration
2 months
Industry and use case
Agriculture, plant-disease detection
DATA
50,000 plant photos
Project duration
2 months
Challenge:
Challenge:
An innovative company in the Agritech sector was building a neural network to automatically identify and classify crop diseases. They collected 50,000 real images of different plants' deseases and asked the Training Data team to annotate this data set. The main challenge was that such data labelling required highly-specialized biological expertise.
Solution:
Solution:
We found and prepared a team of 10 biologists with diverse backgrounds and expertise who could segment the disease areas first and then classify the types of plant illness. Their work led to high-quality dataset annotation that was not possible to do with the internal company's resources in a short timeframe.
Outcomes:
Outcomes:
Timely crop disease detection and proactive mitigation measures
The neural network pilot results demonstrated the potential to increase crop yields by 1-3%
Feedback
Training Data formed a highly-specialized team of biologists for us, and it was the only way we could get the data we needed. Also, they assigned an engagement manager whom we trusted from day one. He was friendly and knowledgeable!
Victor S., Deputy Head of IT Solutions and Innovation Department