Spaces:
Runtime error
Runtime error
Commit
•
e156e53
1
Parent(s):
3f94b52
re-modified code with mediapipe and webcam initialized
Browse files
app.py
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
import os
|
2 |
import cv2
|
3 |
import gradio as gr
|
4 |
import mediapipe as mp
|
@@ -6,14 +6,96 @@ import numpy as np
|
|
6 |
import tensorflow as tf
|
7 |
import tensorflow.lite as tflite
|
8 |
|
9 |
-
# Initialize MediaPipe solutions
|
10 |
-
mp_hands = mp.solutions.hands
|
11 |
-
mp_pose = mp.solutions.pose
|
12 |
-
mp_face_mesh = mp.solutions.face_mesh
|
13 |
|
14 |
-
|
15 |
-
|
16 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
|
18 |
# Get the absolute path to the directory containing app.py
|
19 |
current_dir = os.path.dirname(os.path.abspath(__file__))
|
@@ -25,50 +107,37 @@ model_path = os.path.join(current_dir, model_filename)
|
|
25 |
interpreter = tf.lite.Interpreter(model_path=model_path)
|
26 |
interpreter.allocate_tensors()
|
27 |
|
|
|
28 |
index_to_class = {
|
29 |
"TV": 0, "after": 1, "airplane": 2, "all": 3, "alligator": 4, "animal": 5, "another": 6, "any": 7, "apple": 8, "arm": 9, "aunt": 10, "awake": 11, "backyard": 12, "bad": 13, "balloon": 14, "bath": 15, "because": 16, "bed": 17, "bedroom": 18, "bee": 19, "before": 20, "beside": 21, "better": 22, "bird": 23, "black": 24, "blow": 25, "blue": 26, "boat": 27, "book": 28, "boy": 29, "brother": 30, "brown": 31, "bug": 32, "bye": 33, "callonphone": 34, "can": 35, "car": 36, "carrot": 37, "cat": 38, "cereal": 39, "chair": 40, "cheek": 41, "child": 42, "chin": 43, "chocolate": 44, "clean": 45, "close": 46, "closet": 47, "cloud": 48, "clown": 49, "cow": 50, "cowboy": 51, "cry": 52, "cut": 53, "cute": 54, "dad": 55, "dance": 56, "dirty": 57, "dog": 58, "doll": 59, "donkey": 60, "down": 61, "drawer": 62, "drink": 63, "drop": 64, "dry": 65, "dryer": 66, "duck": 67, "ear": 68, "elephant": 69, "empty": 70, "every": 71, "eye": 72, "face": 73, "fall": 74, "farm": 75, "fast": 76, "feet": 77, "find": 78, "fine": 79, "finger": 80, "finish": 81, "fireman": 82, "first": 83, "fish": 84, "flag": 85, "flower": 86, "food": 87, "for": 88, "frenchfries": 89, "frog": 90, "garbage": 91, "gift": 92, "giraffe": 93, "girl": 94, "give": 95, "glasswindow": 96, "go": 97, "goose": 98, "grandma": 99, "grandpa": 100, "grass": 101, "green": 102, "gum": 103, "hair": 104, "happy": 105, "hat": 106, "hate": 107, "have": 108, "haveto": 109, "head": 110, "hear": 111, "helicopter": 112, "hello": 113, "hen": 114, "hesheit": 115, "hide": 116, "high": 117, "home": 118, "horse": 119, "hot": 120, "hungry": 121, "icecream": 122, "if": 123, "into": 124, "jacket": 125, "jeans": 126, "jump": 127, "kiss": 128, "kitty": 129, "lamp": 130, "later": 131, "like": 132, "lion": 133, "lips": 134, "listen": 135, "look": 136, "loud": 137, "mad": 138, "make": 139, "man": 140, "many": 141, "milk": 142, "minemy": 143, "mitten": 144, "mom": 145, "moon": 146, "morning": 147, "mouse": 148, "mouth": 149, "nap": 150, "napkin": 151, "night": 152, "no": 153, "noisy": 154, "nose": 155, "not": 156, "now": 157, "nuts": 158, "old": 159, "on": 160, "open": 161, "orange": 162, "outside": 163, "owie": 164, "owl": 165, "pajamas": 166, "pen": 167, "pencil": 168, "penny": 169, "person": 170, "pig": 171, "pizza": 172, "please": 173, "police": 174, "pool": 175, "potty": 176, "pretend": 177, "pretty": 178, "puppy": 179, "puzzle": 180, "quiet": 181, "radio": 182, "rain": 183, "read": 184, "red": 185, "refrigerator": 186, "ride": 187, "room": 188, "sad": 189, "same": 190, "say": 191, "scissors": 192, "see": 193, "shhh": 194, "shirt": 195, "shoe": 196, "shower": 197, "sick": 198, "sleep": 199, "sleepy": 200, "smile": 201, "snack": 202, "snow": 203, "stairs": 204, "stay": 205, "sticky": 206, "store": 207, "story": 208, "stuck": 209, "sun": 210, "table": 211, "talk": 212, "taste": 213, "thankyou": 214, "that": 215, "there": 216, "think": 217, "thirsty": 218, "tiger": 219, "time": 220, "tomorrow": 221, "tongue": 222, "tooth": 223, "toothbrush": 224, "touch": 225, "toy": 226, "tree": 227, "uncle": 228, "underwear": 229, "up": 230, "vacuum": 231, "wait": 232, "wake": 233, "water": 234, "wet": 235, "weus": 236, "where": 237, "white": 238, "who": 239, "why": 240, "will": 241, "wolf": 242, "yellow": 243, "yes": 244, "yesterday": 245, "yourself": 246, "yucky": 247, "zebra": 248, "zipper": 249
|
30 |
}
|
31 |
|
32 |
inv_index_to_class = {v: k for k, v in index_to_class.items()}
|
33 |
|
|
|
34 |
|
35 |
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
|
45 |
-
return np.array(combined_landmarks, dtype=np.float32)
|
46 |
|
47 |
-
|
48 |
-
|
49 |
-
if
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
|
58 |
-
if not results.multi_hand_landmarks or not pose_results.pose_landmarks or not face_results.multi_face_landmarks:
|
59 |
-
return None
|
60 |
-
|
61 |
-
hand1_landmarks = results.multi_hand_landmarks[0]
|
62 |
-
if len(results.multi_hand_landmarks) > 1:
|
63 |
-
hand2_landmarks = results.multi_hand_landmarks[1]
|
64 |
-
else:
|
65 |
-
hand2_landmarks = hand1_landmarks
|
66 |
-
|
67 |
-
pose_landmarks = pose_results.pose_landmarks
|
68 |
-
face_landmarks = face_results.multi_face_landmarks[0]
|
69 |
-
lip_landmarks = [face_landmarks.landmark[i] for i in LIPS_IDXS0 - START_IDX]
|
70 |
-
|
71 |
-
return hand1_landmarks, hand2_landmarks, pose_landmarks, lip_landmarks
|
72 |
|
73 |
# Make prediction
|
74 |
def make_prediction(processed_landmarks):
|
@@ -78,47 +147,48 @@ def make_prediction(processed_landmarks):
|
|
78 |
index = outputs[0].argmax()
|
79 |
return index_to_class[index]
|
80 |
|
81 |
-
# Gradio Interface Function
|
82 |
-
def predict_with_webcam(frame):
|
83 |
-
|
84 |
-
# Initialize webcam capture (the default camera, usually index 0)
|
85 |
-
webcam = cv2.VideoCapture(0)
|
86 |
|
87 |
-
|
88 |
-
# Capture a frame from the webcam
|
89 |
-
ret, frame = webcam.read()
|
90 |
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
|
|
96 |
if landmarks is not None:
|
97 |
-
|
98 |
-
|
|
|
|
|
|
|
99 |
print("Prediction:", prediction)
|
|
|
|
|
|
|
|
|
100 |
|
101 |
-
|
102 |
-
|
|
|
|
|
|
|
|
|
103 |
|
104 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
105 |
if cv2.waitKey(1) & 0xFF == ord('q'):
|
106 |
break
|
107 |
|
108 |
-
|
109 |
-
|
110 |
-
cv2.destroyAllWindows()
|
111 |
-
|
112 |
-
landmarks = extract_landmarks(frame)
|
113 |
-
if landmarks is not None:
|
114 |
-
processed_landmarks = preprocess_landmarks(*landmarks)
|
115 |
-
prediction = make_prediction(processed_landmarks)
|
116 |
-
print("Prediction:", prediction)
|
117 |
-
return prediction
|
118 |
-
else:
|
119 |
-
return "Could not detect landmarks. Make sure your webcam is working properly."
|
120 |
|
121 |
-
|
122 |
|
123 |
|
124 |
# Define the Gradio interface with the Webcam input and Text output
|
@@ -133,5 +203,4 @@ webcam_interface = gr.Interface(
|
|
133 |
|
134 |
# Launch the Gradio app with the webcam interface and create a public link
|
135 |
if __name__ == "__main__":
|
136 |
-
webcam_interface.launch()
|
137 |
-
|
|
|
1 |
+
import os
|
2 |
import cv2
|
3 |
import gradio as gr
|
4 |
import mediapipe as mp
|
|
|
6 |
import tensorflow as tf
|
7 |
import tensorflow.lite as tflite
|
8 |
|
|
|
|
|
|
|
|
|
9 |
|
10 |
+
|
11 |
+
# Tensorflow layer to process data in TFLite
|
12 |
+
# Data needs to be processed in the model itself, so we cannot use Python
|
13 |
+
class PreprocessLayer(tf.keras.layers.Layer):
|
14 |
+
def __init__(self):
|
15 |
+
super(PreprocessLayer, self).__init__()
|
16 |
+
|
17 |
+
def pad_edge(self, t, repeats, side):
|
18 |
+
if side == 'LEFT':
|
19 |
+
return tf.concat((tf.repeat(t[:1], repeats=repeats, axis=0), t), axis=0)
|
20 |
+
elif side == 'RIGHT':
|
21 |
+
return tf.concat((t, tf.repeat(t[-1:], repeats=repeats, axis=0)), axis=0)
|
22 |
+
|
23 |
+
@tf.function(
|
24 |
+
input_signature=(tf.TensorSpec(shape=[None, N_ROWS, N_DIMS], dtype=tf.float32),),
|
25 |
+
)
|
26 |
+
def call(self, data0):
|
27 |
+
# Number of Frames in Video
|
28 |
+
N_FRAMES0 = tf.shape(data0)[0]
|
29 |
+
|
30 |
+
# Filter Out Frames With Empty Hand Data
|
31 |
+
frames_hands_nansum = tf.experimental.numpy.nanmean(tf.gather(data0, HAND_IDXS0, axis=1), axis=[1, 2])
|
32 |
+
non_empty_frames_idxs = tf.where(frames_hands_nansum > 0)
|
33 |
+
non_empty_frames_idxs = tf.squeeze(non_empty_frames_idxs, axis=1)
|
34 |
+
data = tf.gather(data0, non_empty_frames_idxs, axis=0)
|
35 |
+
|
36 |
+
# Cast Indices in float32 to be compatible with Tensorflow Lite
|
37 |
+
non_empty_frames_idxs = tf.cast(non_empty_frames_idxs, tf.float32)
|
38 |
+
|
39 |
+
# Number of Frames in Filtered Video
|
40 |
+
N_FRAMES = tf.shape(data)[0]
|
41 |
+
|
42 |
+
# Gather Relevant Landmark Columns
|
43 |
+
data = tf.gather(data, LANDMARK_IDXS0, axis=1)
|
44 |
+
|
45 |
+
# Video fits in INPUT_SIZE
|
46 |
+
if N_FRAMES < INPUT_SIZE:
|
47 |
+
# Pad With -1 to indicate padding
|
48 |
+
non_empty_frames_idxs = tf.pad(non_empty_frames_idxs, [[0, INPUT_SIZE - N_FRAMES]], constant_values=-1)
|
49 |
+
# Pad Data With Zeros
|
50 |
+
data = tf.pad(data, [[0, INPUT_SIZE - N_FRAMES], [0, 0], [0, 0]], constant_values=0)
|
51 |
+
# Fill NaN Values With 0
|
52 |
+
data = tf.where(tf.math.is_nan(data), 0.0, data)
|
53 |
+
return data, non_empty_frames_idxs
|
54 |
+
# Video needs to be downsampled to INPUT_SIZE
|
55 |
+
else:
|
56 |
+
# Repeat
|
57 |
+
if N_FRAMES < INPUT_SIZE ** 2:
|
58 |
+
repeats = tf.math.floordiv(INPUT_SIZE * INPUT_SIZE, N_FRAMES0)
|
59 |
+
data = tf.repeat(data, repeats=repeats, axis=0)
|
60 |
+
non_empty_frames_idxs = tf.repeat(non_empty_frames_idxs, repeats=repeats, axis=0)
|
61 |
+
|
62 |
+
# Pad To Multiple Of Input Size
|
63 |
+
pool_size = tf.math.floordiv(len(data), INPUT_SIZE)
|
64 |
+
if tf.math.mod(len(data), INPUT_SIZE) > 0:
|
65 |
+
pool_size += 1
|
66 |
+
|
67 |
+
if pool_size == 1:
|
68 |
+
pad_size = (pool_size * INPUT_SIZE) - len(data)
|
69 |
+
else:
|
70 |
+
pad_size = (pool_size * INPUT_SIZE) % len(data)
|
71 |
+
|
72 |
+
# Pad Start/End with Start/End value
|
73 |
+
pad_left = tf.math.floordiv(pad_size, 2) + tf.math.floordiv(INPUT_SIZE, 2)
|
74 |
+
pad_right = tf.math.floordiv(pad_size, 2) + tf.math.floordiv(INPUT_SIZE, 2)
|
75 |
+
if tf.math.mod(pad_size, 2) > 0:
|
76 |
+
pad_right += 1
|
77 |
+
|
78 |
+
# Pad By Concatenating Left/Right Edge Values
|
79 |
+
data = self.pad_edge(data, pad_left, 'LEFT')
|
80 |
+
data = self.pad_edge(data, pad_right, 'RIGHT')
|
81 |
+
|
82 |
+
# Pad Non Empty Frame Indices
|
83 |
+
non_empty_frames_idxs = self.pad_edge(non_empty_frames_idxs, pad_left, 'LEFT')
|
84 |
+
non_empty_frames_idxs = self.pad_edge(non_empty_frames_idxs, pad_right, 'RIGHT')
|
85 |
+
|
86 |
+
# Reshape to Mean Pool
|
87 |
+
data = tf.reshape(data, [INPUT_SIZE, -1, N_COLS, N_DIMS])
|
88 |
+
non_empty_frames_idxs = tf.reshape(non_empty_frames_idxs, [INPUT_SIZE, -1])
|
89 |
+
|
90 |
+
# Mean Pool
|
91 |
+
data = tf.experimental.numpy.nanmean(data, axis=1)
|
92 |
+
non_empty_frames_idxs = tf.experimental.numpy.nanmean(non_empty_frames_idxs, axis=1)
|
93 |
+
|
94 |
+
# Fill NaN Values With 0
|
95 |
+
data = tf.where(tf.math.is_nan(data), 0.0, data)
|
96 |
+
|
97 |
+
return data, non_empty_frames_idxs
|
98 |
+
|
99 |
|
100 |
# Get the absolute path to the directory containing app.py
|
101 |
current_dir = os.path.dirname(os.path.abspath(__file__))
|
|
|
107 |
interpreter = tf.lite.Interpreter(model_path=model_path)
|
108 |
interpreter.allocate_tensors()
|
109 |
|
110 |
+
|
111 |
index_to_class = {
|
112 |
"TV": 0, "after": 1, "airplane": 2, "all": 3, "alligator": 4, "animal": 5, "another": 6, "any": 7, "apple": 8, "arm": 9, "aunt": 10, "awake": 11, "backyard": 12, "bad": 13, "balloon": 14, "bath": 15, "because": 16, "bed": 17, "bedroom": 18, "bee": 19, "before": 20, "beside": 21, "better": 22, "bird": 23, "black": 24, "blow": 25, "blue": 26, "boat": 27, "book": 28, "boy": 29, "brother": 30, "brown": 31, "bug": 32, "bye": 33, "callonphone": 34, "can": 35, "car": 36, "carrot": 37, "cat": 38, "cereal": 39, "chair": 40, "cheek": 41, "child": 42, "chin": 43, "chocolate": 44, "clean": 45, "close": 46, "closet": 47, "cloud": 48, "clown": 49, "cow": 50, "cowboy": 51, "cry": 52, "cut": 53, "cute": 54, "dad": 55, "dance": 56, "dirty": 57, "dog": 58, "doll": 59, "donkey": 60, "down": 61, "drawer": 62, "drink": 63, "drop": 64, "dry": 65, "dryer": 66, "duck": 67, "ear": 68, "elephant": 69, "empty": 70, "every": 71, "eye": 72, "face": 73, "fall": 74, "farm": 75, "fast": 76, "feet": 77, "find": 78, "fine": 79, "finger": 80, "finish": 81, "fireman": 82, "first": 83, "fish": 84, "flag": 85, "flower": 86, "food": 87, "for": 88, "frenchfries": 89, "frog": 90, "garbage": 91, "gift": 92, "giraffe": 93, "girl": 94, "give": 95, "glasswindow": 96, "go": 97, "goose": 98, "grandma": 99, "grandpa": 100, "grass": 101, "green": 102, "gum": 103, "hair": 104, "happy": 105, "hat": 106, "hate": 107, "have": 108, "haveto": 109, "head": 110, "hear": 111, "helicopter": 112, "hello": 113, "hen": 114, "hesheit": 115, "hide": 116, "high": 117, "home": 118, "horse": 119, "hot": 120, "hungry": 121, "icecream": 122, "if": 123, "into": 124, "jacket": 125, "jeans": 126, "jump": 127, "kiss": 128, "kitty": 129, "lamp": 130, "later": 131, "like": 132, "lion": 133, "lips": 134, "listen": 135, "look": 136, "loud": 137, "mad": 138, "make": 139, "man": 140, "many": 141, "milk": 142, "minemy": 143, "mitten": 144, "mom": 145, "moon": 146, "morning": 147, "mouse": 148, "mouth": 149, "nap": 150, "napkin": 151, "night": 152, "no": 153, "noisy": 154, "nose": 155, "not": 156, "now": 157, "nuts": 158, "old": 159, "on": 160, "open": 161, "orange": 162, "outside": 163, "owie": 164, "owl": 165, "pajamas": 166, "pen": 167, "pencil": 168, "penny": 169, "person": 170, "pig": 171, "pizza": 172, "please": 173, "police": 174, "pool": 175, "potty": 176, "pretend": 177, "pretty": 178, "puppy": 179, "puzzle": 180, "quiet": 181, "radio": 182, "rain": 183, "read": 184, "red": 185, "refrigerator": 186, "ride": 187, "room": 188, "sad": 189, "same": 190, "say": 191, "scissors": 192, "see": 193, "shhh": 194, "shirt": 195, "shoe": 196, "shower": 197, "sick": 198, "sleep": 199, "sleepy": 200, "smile": 201, "snack": 202, "snow": 203, "stairs": 204, "stay": 205, "sticky": 206, "store": 207, "story": 208, "stuck": 209, "sun": 210, "table": 211, "talk": 212, "taste": 213, "thankyou": 214, "that": 215, "there": 216, "think": 217, "thirsty": 218, "tiger": 219, "time": 220, "tomorrow": 221, "tongue": 222, "tooth": 223, "toothbrush": 224, "touch": 225, "toy": 226, "tree": 227, "uncle": 228, "underwear": 229, "up": 230, "vacuum": 231, "wait": 232, "wake": 233, "water": 234, "wet": 235, "weus": 236, "where": 237, "white": 238, "who": 239, "why": 240, "will": 241, "wolf": 242, "yellow": 243, "yes": 244, "yesterday": 245, "yourself": 246, "yucky": 247, "zebra": 248, "zipper": 249
|
113 |
}
|
114 |
|
115 |
inv_index_to_class = {v: k for k, v in index_to_class.items()}
|
116 |
|
117 |
+
mp_holistic = mp.solutions.holistic
|
118 |
|
119 |
|
120 |
+
def mediapipe_detection(image, model):
|
121 |
+
# COLOR CONVERSION BGR 2 RGB
|
122 |
+
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
|
123 |
+
image.flags.writeable = False # Image is no longer writeable
|
124 |
+
results = model.process(image) # Make prediction
|
125 |
+
image.flags.writeable = True # Image is now writeable
|
126 |
+
image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR) # COLOR COVERSION RGB 2 BGR
|
127 |
+
return image, results
|
128 |
|
|
|
129 |
|
130 |
+
def extract_keypoints(results):
|
131 |
+
lh = np.array([[res.x, res.y, res.z] for res in results.left_hand_landmarks.landmark]).flatten(
|
132 |
+
) if results.left_hand_landmarks else np.zeros(21*3)
|
133 |
+
rh = np.array([[res.x, res.y, res.z] for res in results.right_hand_landmarks.landmark]).flatten(
|
134 |
+
) if results.right_hand_landmarks else np.zeros(21*3)
|
135 |
+
pose = np.array([[res.x, res.y, res.z, res.visibility] for res in results.pose_landmarks.landmark]).flatten(
|
136 |
+
) if results.pose_landmarks else np.zeros(33*4)
|
137 |
+
face = np.array([[res.x, res.y, res.z] for res in results.face_landmarks.landmark]).flatten(
|
138 |
+
) if results.face_landmarks else np.zeros(468*3)
|
139 |
+
return np.concatenate([lh, rh, pose, face])
|
140 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
141 |
|
142 |
# Make prediction
|
143 |
def make_prediction(processed_landmarks):
|
|
|
147 |
index = outputs[0].argmax()
|
148 |
return index_to_class[index]
|
149 |
|
|
|
|
|
|
|
|
|
|
|
150 |
|
151 |
+
# ... (previous code)
|
|
|
|
|
152 |
|
153 |
+
def predict_with_webcam(frame):
|
154 |
+
with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:
|
155 |
+
# Make detections using mediapipe
|
156 |
+
image, results = mediapipe_detection(frame, holistic)
|
157 |
+
print(results)
|
158 |
+
landmarks = extract_keypoints(results)
|
159 |
if landmarks is not None:
|
160 |
+
# Initialize PreprocessLayer
|
161 |
+
preprocess_layer = PreprocessLayer()
|
162 |
+
# Call the PreprocessLayer to preprocess the landmarks
|
163 |
+
processed_landmarks, _ = preprocess_layer.call(landmarks)
|
164 |
+
prediction = make_prediction(processed_landmarks) # Pass the preprocessed landmarks to make_prediction
|
165 |
print("Prediction:", prediction)
|
166 |
+
return prediction
|
167 |
+
else:
|
168 |
+
return "Could not detect landmarks. Make sure your webcam is working properly."
|
169 |
+
|
170 |
|
171 |
+
cap = cv2.VideoCapture(0)
|
172 |
+
# Set mediapipe model
|
173 |
+
with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:
|
174 |
+
while cap.isOpened():
|
175 |
+
# Read feed
|
176 |
+
ret, frame = cap.read()
|
177 |
|
178 |
+
# Make predictions
|
179 |
+
prediction = predict_with_webcam(frame)
|
180 |
+
|
181 |
+
# Display the frame with the prediction
|
182 |
+
cv2.putText(frame, prediction, (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
|
183 |
+
cv2.imshow('Webcam Landmark Prediction', frame)
|
184 |
+
|
185 |
+
# Exit the loop when 'q' key is pressed
|
186 |
if cv2.waitKey(1) & 0xFF == ord('q'):
|
187 |
break
|
188 |
|
189 |
+
cap.release()
|
190 |
+
cv2.destroyAllWindows()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
191 |
|
|
|
192 |
|
193 |
|
194 |
# Define the Gradio interface with the Webcam input and Text output
|
|
|
203 |
|
204 |
# Launch the Gradio app with the webcam interface and create a public link
|
205 |
if __name__ == "__main__":
|
206 |
+
webcam_interface.launch()
|
|