search

OlafenwaMoses / ImageAI

A python library built to empower developers to build applications and systems with self-contained Computer Vision capabilities
https://www.genxr.co/#products
MIT License
8.56k stars 2.19k forks source link

Can I use Custom Detection Model Training to identify images in the following TIF format without converting to JPG images? #392

Closed HaoNiu1997 closed 4 years ago

HaoNiu1997 commented 4 years ago

Sir,I want to train some grayscale pictures taken by electron microscopy. The pictures are as follows. The left and right images are a picture. The left side is opened with special software, and the picture on the right is directly displayed. Is the program reading the right picture when it is being trained? Can I accurately identify the image I want to recognize when I make a prediction? Can I identify when there are multiple categories on one image? 图片

AbimbolaOO commented 4 years ago

No, you can't do so. Firstly, you need to understand that TIF is a very high-resolution image format used for print at the press. Most computer vision algorithm work with JPEG and PNG because they don't need to metadata store by TIF format.