Spaces:
Build error
Build error
File size: 3,100 Bytes
1008a6e 79b1351 1008a6e 79b1351 1008a6e 79b1351 1008a6e 79b1351 1008a6e 79b1351 1008a6e 79b1351 1008a6e |
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 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 |
import gradio as gr
import numpy as np
import cv2
# model load
cfg = r'volleyball_test.cfg'
weights = r'volleyball_final.weights'
net = cv2.dnn.readNetFromDarknet(cfg, weights)
# classes
classes = []
with open("classes.names", 'r') as f:
classes = f.read().splitlines()
def predict_img(img_bgr):
# img_bgr = inp.astype('uint8')[...,::-1]
img = cv2.resize(img_bgr, (700, 700))
height, width, channels = img.shape
# Convert image into blob and load it on model
blob = cv2.dnn.blobFromImage(
img, 1/255, (height, width), (0, 0, 0), swapRB=True, crop=False)
net.setInput(blob)
# Getting all the three detection layers of yolo
output_layers_names = net.getUnconnectedOutLayersNames()
# print(output_layers_names)
layersOutputs = net.forward(output_layers_names)
# print(layersOutputs)
# Finding the y-vector and minimum no.of bounding box
confthreshold = 0.5
boxes = []
confidences = []
class_ids = []
for output in layersOutputs:
for detection in output:
scores = detection[5:]
class_id = np.argmax(scores)
confidence = scores[class_id]
if confidence > confthreshold:
center_x = int(detection[0]*width)
center_y = int(detection[1]*height)
w = int(detection[2]*width)
h = int(detection[3]*height)
x = int(center_x - w/2)
y = int(center_y - h/2)
boxes.append([x, y, w, h])
confidences.append(float(confidence))
class_ids.append(class_id)
# Applying Non max Suppression for removing unwanted multiple bounding boxes
indexes = cv2.dnn.NMSBoxes(
boxes, confidences, confthreshold, nms_threshold=0.3)
for i in indexes:
box = boxes[i]
x, y, w, h = box
cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 0), 2)
conf_value = str(round(confidences[i], 2))
label = str(classes[class_ids[i]])
cv2.putText(img, label + " " + conf_value, (x, y-10),
cv2.FONT_HERSHEY_COMPLEX, 1, (0, 255, 0), 2)
# return the images
# cv2.imwrite("out.jpg", img)
return img
def predict(inp):
vidcap = cv2.VideoCapture(inp)
fps = vidcap.get(cv2.CAP_PROP_FPS)
outcap = cv2.VideoWriter('outpy.mp4',cv2.VideoWriter_fourcc(*'MP4V'), fps, (700, 700))
success,image = vidcap.read()
count = 0
while success:
img = predict_img(image)
outcap.write(img)
success,image = vidcap.read()
print('Read a new frame: ', success)
count += 1
return "./outpy.mp4"
gr.Interface(
fn=predict,
inputs=[
gr.inputs.Video() # you can have many inputs
],
outputs=[
gr.inputs.Video() # you can have many outputs
],
title="Volley classification and detection",
description="This project use a yolov3 and pretrained model from [this](https://github.com/lalchhabi/Volleyball_Position_Detection_System) project",
examples=[
"test.mp4",
]
).launch()
|