app.py

import gradio as gr
from PIL import Image
import torch
from transformers import SamModel, SamProcessor
import numpy as np
import matplotlib
matplotlib.use("Agg")
print("Matplotlib version:", matplotlib.__version__)
import matplotlib.pyplot as plt

#Load the SAM model and processor
model = SamModel.from_pretrained("Zigeng/SlimSAM-uniform-77")
processor = SamProcessor.from_pretrained("Zigeng/SlimSAM-uniform-77")

#Global variable to store input points
input_points = []

#Helper functions
def show_mask(mask, ax, random_color=False):
    if random_color:
        color = np.concatenate([np.random.random(3),
                               np.array([0.6])],
                              axis=0)
    else:
        color = np.array([30/255, 144/255, 255/255, 0.6])
    h, w = mask.shape[-2:]
    mask_image = mask.reshape(h,w,1)*color.reshape(1,1,-1)
    ax.imshow(mask_image)

#Function to get pixel coordinates
def get_pixel_coordinates(image, evt:gr.SelectData):
    global input_points
    x, y = evt.index[0], evt.index[1]
    input_points = [[[x, y]]]
    return perform_prediction(image)

#Function to perform SAM model prediction
def perform_prediction(image):
    global input_points
    #Preprocess the image
    inputs = processor(images=image, input_points=input_points, return_tensors="pt")
    #Perform prediction
    with torch.no_grad():
        outputs = model(**inputs)
    iou = outputs.iou_scores
    max_iou_index = torch.argmax(iou)

    #Post-process the masks
    predicted_masks = processor.image_processor.post_process_masks(
        outputs.pred_masks,
        inputs['original_sizes'],
        inputs['reshaped_input_sizes']
    )
    predicted_mask = predicted_masks[0]

    #Display the mask on the image
    mask_image = show_mask_on_image(image, predicted_mask[:, max_iou_index], return_image=True)
    return mask_image

# Function to overlay mask on the image
def show_mask_on_image(raw_image, mask, return_image=False):
    if not isinstance(mask, torch.Tensor):
        mask = torch.Tensor(mask)

    if len(mask.shape) == 4:
        mask = mask.squeeze()

    fig, axes = plt.subplots(1, 1, figsize=(15, 15))

    mask = mask.cpu().detach()
    axes.imshow(np.array(raw_image))
    show_mask(mask, axes)
    axes.axis("off")
    plt.show()

    if return_image:
        fig = plt.gcf()
        fig.canvas.draw()
        # Convert plot to image
        img = np.frombuffer(fig.canvas.tostring_argb(), dtype=np.uint8)
        img = img.reshape(fig.canvas.get_width_height()[::-1] + (3,))
        img = Image.fromarray(img)
        plt.close(fig)
        return img

#Create the Gradio interface
with gr.Blocks() as demo:
    with gr.Row():
        img = gr.Image(type="pil", label="Input Image", height=400, width=600)
        output_image = gr.Image(label="Masked Image")
    img.select(get_pixel_coordinates, inputs=[img], outputs=[output_image])

demo.launch(share=False)