1 of 5

Saad Hassan, Oliver Alonzo, Abraham Glasser, Matt Huenerfauth

SLRTP’20 at ECCV

Effect of Ranking and Precision of Results on Users' Satisfaction with Search-by-Video Sign-Language Dictionaries

| 1

2 of 5

Prior work

Sign recognition technology is being used to create tools to let people search for sign in an ASL dictionary by performing what they want into a webcam.�
We do not know how does performance affect user judgement?

| 2

Procedure:

Wizard-of-oz Study

Stimulus

Webcam Recording

Results Page

Questions

| 3

What we learnt?

When using “top-k” metrics, the range between 10 and 20 is important.
In addition to considering the position of the desired word, the precision of the overall list matters.

Provide guidance for researchers studying sign-language dictionary systems or the underlying technologies for identifying matches.

| 4

Thanks

	Matt Huenerfauth Professor Rochester Institute of Technology Golisano College of Computing and Information School of Information - The iSchool Email: matt.huenerfauth@rit.edu

	Oliver Alonzo Ph.D. Student Rochester Institute of Technology Computing and Information Sciences Ph.D. Program Email: oa7652@rit.edu

	Abraham Glasser Ph.D. Student Rochester Institute of Technology Computing and Information Sciences Ph.D. Program Email: atg2036@rit.edu

Acknowledgements
This material is based upon work supported by the National Science Foundation under award No. 1763569. We are grateful for the contribution of Aakash Maddi in the collection of data from participants for this work.

	Saad Hassan Ph.D. Student Rochester Institute of Technology Computing and Information Sciences Ph.D. Program Email: sh2513@rit.edu

| 5