app.py

import openvino as ov
import gradio as gr
import yaml
import cv2 
import numpy as np
from ultralytics.utils.plotting import colors
core = ov.Core()

model = core.read_model(model = "models/best.xml")
compiled_model = core.compile_model(model = model, device_name = "AUTO")

input_layer = compiled_model.input(0)
output_layer = compiled_model.output(0)

with open('models/metadata.yaml') as info:
    info_dict = yaml.load(info, Loader=yaml.Loader)

labels = info_dict['names']

def prepare_data(image, input_layer):
    input_w, input_h = input_layer.shape[2], input_layer.shape[3]
 
    input_image = cv2.resize(image, (input_w, input_h))
    input_image = cv2.cvtColor(input_image, cv2.COLOR_BGR2RGB)
    input_image = input_image/255
    input_image = input_image.transpose(2,0,1)
    input_image = np.expand_dims(input_image, 0)
 
    return input_image

def evaluate(output, conf_threshold):
    
    boxes = []
    scores = []
    label_key = []
    label_index = 0
    
    for class_ in output[0][4:]:
  
        for index in range (len(class_)):
            confidence = class_[index]

            if  confidence > conf_threshold:

                xcen = output[0][0][index]
                ycen = output[0][1][index]
                w = output[0][2][index]
                h = output[0][3][index]

                xmin = int(xcen - (w/2))
                xmax = int(xcen + (w/2))
                ymin = int(ycen - (h/2))
                ymax = int(ycen + (h/2))

                box = (xmin, ymin, xmax, ymax)
                boxes.append(box)
                scores.append(confidence)

                label_key.append(label_index)
  
        label_index += 1 
        
    boxes = np.array(boxes)
    scores = np.array(scores)
    
    return boxes, scores, label_key

def non_max_suppression(boxes, scores, threshold):	
    assert boxes.shape[0] == scores.shape[0]
    # bottom-left origin
    ys1 = boxes[:, 0]
    xs1 = boxes[:, 1]
    # top-right target
    ys2 = boxes[:, 2]
    xs2 = boxes[:, 3]
    # box coordinate ranges are inclusive-inclusive
    areas = (ys2 - ys1) * (xs2 - xs1)
    scores_indexes = scores.argsort().tolist()
    boxes_keep_index = []
    while len(scores_indexes):
        index = scores_indexes.pop()
        boxes_keep_index.append(index)
        if not len(scores_indexes):
            break
        ious = compute_iou(boxes[index], boxes[scores_indexes], areas[index],
                           areas[scores_indexes])
        filtered_indexes = set((ious > threshold).nonzero()[0])
        # if there are no more scores_index
        # then we should pop it
        scores_indexes = [
            v for (i, v) in enumerate(scores_indexes)
            if i not in filtered_indexes
        ]
    return np.array(boxes_keep_index)


def compute_iou(box, boxes, box_area, boxes_area):
    # this is the iou of the box against all other boxes
    assert boxes.shape[0] == boxes_area.shape[0]
    # get all the origin-ys
    # push up all the lower origin-xs, while keeping the higher origin-xs
    ys1 = np.maximum(box[0], boxes[:, 0])
    # get all the origin-xs
    # push right all the lower origin-xs, while keeping higher origin-xs
    xs1 = np.maximum(box[1], boxes[:, 1])
    # get all the target-ys
    # pull down all the higher target-ys, while keeping lower origin-ys
    ys2 = np.minimum(box[2], boxes[:, 2])
    # get all the target-xs
    # pull left all the higher target-xs, while keeping lower target-xs
    xs2 = np.minimum(box[3], boxes[:, 3])
    # each intersection area is calculated by the
    # pulled target-x minus the pushed origin-x
    # multiplying
    # pulled target-y minus the pushed origin-y
    # we ignore areas where the intersection side would be negative
    # this is done by using maxing the side length by 0
    intersections = np.maximum(ys2 - ys1, 0) * np.maximum(xs2 - xs1, 0)
    # each union is then the box area
    # added to each other box area minusing their intersection calculated above
    unions = box_area + boxes_area - intersections
    # element wise division
    # if the intersection is 0, then their ratio is 0
    ious = intersections / unions
    return ious
    
def visualize(image, nms_output, boxes, label_key,scores, conf_threshold):
    image_h, image_w, c = image.shape
    input_w, input_h = input_layer.shape[2], input_layer.shape[3]

    for i in nms_output:
        xmin, ymin, xmax, ymax = boxes[i]
 
        xmin = int(xmin*image_w/input_w)
        xmax = int(xmax*image_w/input_w)
        ymin = int(ymin*image_h/input_h)
        ymax = int(ymax*image_h/input_h)
 
        label = label_key[i]
        color = colors(label)
        cv2.rectangle(image, (xmin, ymin), (xmax, ymax), color, 1)
        font = cv2.FONT_HERSHEY_SIMPLEX
        text = str(int(scores[i]*100)) + "%" + labels[label]
        font_scale= (image_w/1000)
        label_width, label_height = cv2.getTextSize(text, font,font_scale, 1)[0]
        cv2.rectangle(image, (xmin, ymin-label_height), (xmin + label_width, ymin), color, -1)
        
        cv2.putText(image, text, (xmin+2, ymin), font, font_scale, (255,255,255), 1, cv2.LINE_AA)    
    return image

def predict_image(image, conf_threshold = .4):

    if image is not None: 
        image_RGB = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
        input_image = prepare_data(image_RGB, input_layer)
        output = compiled_model([input_image])[output_layer]
        boxes, scores, label_key = evaluate(output, conf_threshold)
     
        if len(boxes):
            nms_output = non_max_suppression(boxes, scores, conf_threshold)
      
            visualized_image = visualize(image_RGB, nms_output, boxes, label_key,scores, conf_threshold)
            visualized_image = cv2.cvtColor(visualized_image, cv2.COLOR_BGR2RGB)
            return visualized_image
        else:
            return image

image_interface = gr.Interface(
    fn = predict_image,
    inputs = [gr.Image(label="Upload Image"),
              gr.Slider(minimum=0.05, maximum = 1, value = .4, label = "Confidence")
             ],
    outputs = gr.Image(label="Results"),
    title = "AI Kickboard Safety Project", 
    description = "Upload images for Inference on YOLOv8.",
    live = True
)


if __name__=="__main__":
    image_interface.launch(share=True)