app.py

import gradio as gr
import numpy as np
import cv2
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

def perform_kmeans_segmentation(image, number_of_clusters):
    # Reshaping the image into a 2D array of pixels and 3 color values (RGB)
    pixel_vals = image.reshape((-1,3))
    # Convert to float type
    pixel_vals = np.float32(pixel_vals)

    criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 100, 0.85)
    k = number_of_clusters
    retval, labels, centers = cv2.kmeans(pixel_vals, k, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)

    # convert data into 8-bit values
    centers = np.uint8(centers)
    segmented_data = centers[labels.flatten()]
    
    # reshape data into the original image dimensions
    segmented_image = segmented_data.reshape((image.shape))
    return segmented_image, centers, labels


def graph(image, centers, labels, display_centres):
    r, g, b = cv2.split(image)
    r = r.flatten()
    g = g.flatten()
    b = b.flatten()
    #plotting 
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    ax.scatter(r, g, b, s=10, c=labels, marker="o", label='image')
    if display_centres:
        ax.scatter(centers[:,0], centers[:,1], centers[:,2], s=50, c='r', marker="o", label='cluster centers')
    return fig


def resize_image(image):
    x = np.shape(image)[0]
    y = np.shape(image)[1]
    if (x <= 800) and (y <= 800):
        return image

    if x >= y:
        r = 800 / x
        dim = (int(y * r), 800)
    else:
        r = 800 / y
        dim = (800, int(x * r))

    res = cv2.resize(image, dim, interpolation=cv2.INTER_AREA)
    return res


def kmeans_image(number_of_clusters, input_image, display_cluster_centres):
    image = cv2.cvtColor(input_image, cv2.COLOR_BGR2RGB)
    image = resize_image(image)

    segmented_image, cluster_centres, labels = perform_kmeans_segmentation(image, number_of_clusters)
    plot = graph(image, cluster_centres, labels, display_cluster_centres)

    return gr.Image(value=segmented_image), plot


iface = gr.Interface(
    fn=kmeans_image, 
    inputs=[gr.Number(minimum=2, maximum=10, value=2, label="Number of Clusters", info="Number of clusters to segment the image into. Enter a value between 2 and 10."), 
            gr.Image(type="numpy", label="Image to Segment"),
            gr.Checkbox(label="display cluster centers")], 
    outputs=["image", gr.Plot(visible=True, label="Image plotted in 3D space")],
    examples=[
        [2, "images/simple_dots.png", False],
        [3, "images/purple_flower.jpg", False],
        [3, "images/rocky_coast.jpg", False],
        [2, "images/simple_flowers.jpg", False],
    ])
iface.launch()