File size: 2,164 Bytes
f645a79 9bfc5a2 f645a79 9bfc5a2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 |
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import gradio as gr
model = SentenceTransformer("sentence-transformers/clip-ViT-B-16")
def predict(im1, im2):
image_embs = model.encode([im1, im2])
similarities = cosine_similarity(image_embs)
sim = similarities[0][1]
threshold = 0.65
if sim > threshold:
return sim, "SAME PERSON, UNLOCK PHONE"
else:
return sim, "DIFFERENT PEOPLE, DON'T UNLOCK"
with gr.Blocks() as demo:
gr.Markdown("Based on two images, the goal is to recognize the similarities/differences between facial images and determine whether or not to unlock a phone based on a cosine similarity score.")
with gr.Tab("Image"):
with gr.Row():
with gr.Column():
img_inputs = [gr.Image(type="pil", source="upload"),
gr.Image(type="pil", source="upload")]
examples = gr.Examples([["https://live.staticflickr.com/2883/33785597726_47880fa539_b.jpg","https://live.staticflickr.com/65535/49086637987_f7622c3345.jpg"],
["https://live.staticflickr.com/3423/3197571945_123937185f_b.jpg", "https://live.staticflickr.com/7259/7001667239_11cece02c8_b.jpg"],
["https://live.staticflickr.com/4015/4334237247_08af133b4b_b.jpg", "https://live.staticflickr.com/3701/9364116426_87b8918e9d_b.jpg"]],
inputs=img_inputs)
btn = gr.Button("Run")
with gr.Column():
btn.click(fn=predict,
inputs=img_inputs,
outputs=[gr.Number(label="Similarity"),
gr.Textbox(label="Message")],
)
with gr.Tab("Webcam"):
with gr.Row():
with gr.Column():
img_inputs = [gr.Image(type="pil", source="webcam"),
gr.Image(type="pil", source="webcam")]
btn = gr.Button("Run")
with gr.Column():
btn.click(fn=predict,
inputs=img_inputs,
outputs=[gr.Number(label="Similarity"),
gr.Textbox(label="Message")],
)
demo.launch(debug=True)
|