Swangya Saurav

Document Type

Honors Thesis


Recognizing object in an image is a complicated task and combines a plethora of variables, which deviates substantially with a small deviation in the environmental conditions. Taking into account all of those variables and programming an object detector to work in multiple conditions is a difficult task. However, using neural networks makes accounting for these variables a lot easier, as it eliminates the requirement of human involvement in setting up those variables. Further, developments in new neural network architectures like Faster RCNN and Mobilenet have made the prediction using neural networks more efficient and accurate than ever. These advancements have enabled us to create applications that require less computational power, yet can produce an accurate result. This technology was used to create a measuring tool that can be used by anyone to easily determine the length of a fish. The neural network for this project was trained to recognize a fish and a US quarter coin in an image, and upon detection of those two objects, the code compared the length of fish to the length of a coin to approximately deduce the length of the fish in imperial units.

Publication Date






To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.