We are building an application that analyses images for repairs diagnosis. GPT vision consistently makes mistakes with boiler pressure gauges. It describes a high pressure fault when the gauge shows low pressure. It also gets appliance error codes incorrect. We have tried combining with OpenCV libraries but it still makes mistakes. What techniques are others using to overcome the limitations?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| GPT4-Vision API cant read radial dials | 0 | 372 | December 8, 2023 | |
| Image recognition of brands clothes and mistakes | 2 | 100 | August 4, 2024 | |
| How to solve the problem that GPT-API cannot read text using OCR? | 19 | 2146 | July 10, 2024 | |
| Vision API flips numbers on extracting text from image | 3 | 1010 | December 13, 2023 | |
| My GPT is not reading images well enough | 3 | 129 | August 13, 2024 |