SignBridge Server

Server/Backend for SignBridge App

• Use Docker for easy deployment.
• Needs python3.8, no other version works becuase it is dependent on torch=1.1.0 which is used by text2sign
• more customizations or control on API is possible

Docs

• Base URL

https://bipinkrish-signbridge.hf.space

1. Text to Gloss

Takes a spoken english text as input and returns simplified gloss of it.

• Link

<base_url>/text2gloss

• Params

text

• Example

https://bipinkrish-signbridge.hf.space/text2gloss?text=membership%20of%20parliament%20see%20minutes

{"gloss": "parliament see minute"}

2. Gloss to Sign

Takes a gloss text as input and returns fsw string which represents ham notation.

• Link

<base_url>/gloss2sign

• Params

gloss

• Example

https://bipinkrish-signbridge.hf.space/gloss2sign?gloss=today

{"sign": "M530x518S19a30500x482S19a38465x481S22f04509x506S22f14467x504"}

3. Sign to Image

Takes a fsw string as input and returns image of ham notation.

• Link

<base_url>/sign2img

• Params

sign
line_color (Optional)

• Example

https://bipinkrish-signbridge.hf.space/sign2img?sign=M530x518S19a30500x482S19a38465x481S22f04509x506S22f14467x504&line_color=10,23,122,255

{"img": img_base64}

4. Gloss to Pose

Takes a gloss text as input and returns combined pose vile visalized as image and words which are considerd.

• Link

<base_url>/gloss2pose

• Params

gloss

• Example

https://bipinkrish-signbridge.hf.space/gloss2pose?gloss=1 2 3

{"img": img_base64, "words": ["one", "two", "three"]}