We are seeking a post-doctoral fellow in Computer Vision at the California Institute of Technology. The initial appointment is for one year renewable yearly for up to three years. 

 

The post-doctoral fellow will be the lead scientist for a project that applies Computer Vision to Ecology and Conservation. The goal of the project is counting salmon traveling upstream in rivers of the Pacific Northwest () from underwater ultrasound video collected at thirty river locations across  Alaska, Washington, Oregon and California.

 

There are two main technical challenges where the project pushes the envelope of current computer vision and machine learning capabilities. The first is self-initialization / domain adaptation -- the system will need to count fish accurately at new locations beyond the training sites and in the face of variations in the environmental conditions.  The second is species classification. The long-term scope of these challenges includes all applications of Computer Vision to Ecology where animals are to be detected, counted, and classified with minimal prior training of the system.

 

The post-doc will report to Professors Pietro Perona and Georgia Gkioxari and will be responsible for the scientific and engineering success of the project. The post-doc will co-advise two graduate students and will manage a software engineer. The project will also involve a team at MIT led by Sara Beery and a team at U. Mass Amherst led by Grant van Horn.