FRAGMENTED LETTERS

[sarithapillai8]

FRAGMENTED LETTERS
https://development.d29wtym67es164.amplifyapp.com/
w/ voice recognition

Options (chosen by researcher through the Lamp portal)
Starting Fragmentation: 30%, 40%, 50%

Images: 26 letters of the alphabet. Black letters on white background. Black regions (letters) split into pixels. Fragmentation in 5% increments 5%-95%.

Fragmentation: 50% fragmentation = a random 50% of the black pixels flipped to white.

Protocol:

1. Randomly choose 1 letter from the English alphabet
2. Fragment per starting fragmentation choice (options)
3. If correct, fragment next image (chosen randomly) 10% more
4. If incorrect, fragment next image 5% less
5. Stop after 2 errors.

Scoring:
Sequence: e.g. (30% correct, 40% correct, 50% incorrect, 45% incorrect)
Best Correct Fragmentation % (40%)

[evinkalengaden]

Hi @kijote0

[Customer Comments_01152025]: In addition, I recognize that the Fragmented Letters task is still going through the QA process. However, when trying it out, the voice recording could not pick up on any letters that I was saying out loud. This may require some fine tuning during the QA testing period.

[ZCO Comments_01212025]: In Fragmented letters activity, Can we ask the user to say a word “Letter” before the alphabet to be identified. We are facing an issue with the system on recognizing an alphabet by vocal. Voice of single alphabet is not getting detected in most of the cases. So we can provide an instruction to the users mentioning the same.

Let me know your comments.

Thanks,
Evin K Alengaden

[kijote0]

Hi @evinkalengaden, that sounds reasonable to me