The problem SignPal solves

There are 466M hearing-impaired people worldwide, which is about 6% of the entire world population, and there is a severe lack of inclusivity.

What do we bring to you?:
1 . Inclusive Communication: SIGNPAL will empower the deaf community by providing real-time translation of audio content, making communication more inclusive and accessible.

2 . Versatile Audio Input: The tool is designed to handle various audio sources, ensuring comprehensive coverage for different types of media.

3 . Digital Avatar: Our unique digital avatar will accurately perform ISL translations, providing a visual representation of spoken content.

4 . Real-Time Translation: With advanced AI and machine learning algorithms, SIGNPAL will deliver instantaneous translations, ensuring that users do not miss out on important conversations or information.

This system has better accuracy and less memory-utilization than existing ones,
Moreover, the system is capable of generating the output within 1s per conversion by minimizing the operational processing time by nearly 69%.

Challenges we ran into

Connecting all the endpoints together was a mammoth task. The codes required working with alot of APIs , downloading resource intensive SigML , HamNoSys, and blender files for our avatar generation. Our model showed alot of latency on cpu while it would have worked seamlessly had we had a GPU. Making the website was a huge task in itself from managing the blender apis to the Django backend.