Update app.py
Browse files
app.py
CHANGED
|
@@ -4,7 +4,7 @@ import numpy as np
|
|
| 4 |
import torch
|
| 5 |
from ultralytics import YOLO
|
| 6 |
import gradio as gr
|
| 7 |
-
from scipy.interpolate import interp1d
|
| 8 |
import uuid
|
| 9 |
import os
|
| 10 |
|
|
@@ -16,9 +16,9 @@ STUMPS_WIDTH = 0.2286 # meters (width of stumps)
|
|
| 16 |
FRAME_RATE = 20 # Input video frame rate
|
| 17 |
SLOW_MOTION_FACTOR = 2 # Reduced for faster output
|
| 18 |
CONF_THRESHOLD = 0.25 # Confidence threshold for detection
|
| 19 |
-
PITCH_ZONE_Y = 0.
|
| 20 |
-
IMPACT_ZONE_Y = 0.
|
| 21 |
-
IMPACT_DELTA_Y =
|
| 22 |
STUMPS_HEIGHT = 0.711 # meters (height of stumps)
|
| 23 |
|
| 24 |
def process_video(video_path):
|
|
@@ -61,7 +61,7 @@ def estimate_trajectory(ball_positions, detection_frames, frames):
|
|
| 61 |
return None, None, None, None, None, None, "Error: Fewer than 2 valid single-ball detections for trajectory"
|
| 62 |
frame_height = frames[0].shape[0]
|
| 63 |
|
| 64 |
-
# Filter to unique positions
|
| 65 |
unique_positions = [ball_positions[0]]
|
| 66 |
for pos in ball_positions[1:]:
|
| 67 |
if abs(pos[0] - unique_positions[-1][0]) > 10 or abs(pos[1] - unique_positions[-1][1]) > 10:
|
|
@@ -70,6 +70,10 @@ def estimate_trajectory(ball_positions, detection_frames, frames):
|
|
| 70 |
y_coords = [pos[1] for pos in unique_positions]
|
| 71 |
times = np.array([i / FRAME_RATE for i in range(len(unique_positions))])
|
| 72 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
pitch_idx = 0
|
| 74 |
for i, y in enumerate(y_coords):
|
| 75 |
if y > frame_height * PITCH_ZONE_Y:
|
|
@@ -94,12 +98,12 @@ def estimate_trajectory(ball_positions, detection_frames, frames):
|
|
| 94 |
times = times[:impact_idx + 1]
|
| 95 |
|
| 96 |
try:
|
| 97 |
-
fx = interp1d(times,
|
| 98 |
-
fy = interp1d(times,
|
| 99 |
except Exception as e:
|
| 100 |
return None, None, None, None, None, None, f"Error in trajectory interpolation: {str(e)}"
|
| 101 |
|
| 102 |
-
vis_trajectory = list(zip(
|
| 103 |
t_full = np.linspace(times[0], times[-1] + 0.5, len(times) + 5)
|
| 104 |
x_full = fx(t_full)
|
| 105 |
y_full = fy(t_full)
|
|
@@ -118,9 +122,9 @@ def lbw_decision(ball_positions, full_trajectory, frames, pitch_point, impact_po
|
|
| 118 |
|
| 119 |
frame_height, frame_width = frames[0].shape[:2]
|
| 120 |
stumps_x = frame_width / 2
|
| 121 |
-
stumps_y = frame_height * 0.
|
| 122 |
stumps_width_pixels = frame_width * (STUMPS_WIDTH / 3.0)
|
| 123 |
-
batsman_area_y = frame_height * 0.
|
| 124 |
|
| 125 |
pitch_x, pitch_y = pitch_point
|
| 126 |
impact_x, impact_y = impact_point
|
|
@@ -150,7 +154,7 @@ def generate_slow_motion(frames, vis_trajectory, pitch_point, pitch_frame, impac
|
|
| 150 |
return None
|
| 151 |
frame_height, frame_width = frames[0].shape[:2]
|
| 152 |
stumps_x = frame_width / 2
|
| 153 |
-
stumps_y = frame_height * 0.
|
| 154 |
stumps_width_pixels = frame_width * (STUMPS_WIDTH / 3.0)
|
| 155 |
stumps_height_pixels = frame_height * (STUMPS_HEIGHT / 3.0)
|
| 156 |
|
|
@@ -160,33 +164,38 @@ def generate_slow_motion(frames, vis_trajectory, pitch_point, pitch_frame, impac
|
|
| 160 |
trajectory_points = np.array(vis_trajectory, dtype=np.int32).reshape((-1, 1, 2))
|
| 161 |
|
| 162 |
for i, frame in enumerate(frames):
|
| 163 |
-
# Draw stumps
|
| 164 |
cv2.line(frame, (int(stumps_x - stumps_width_pixels / 2), int(stumps_y)),
|
| 165 |
(int(stumps_x + stumps_width_pixels / 2), int(stumps_y)), (255, 255, 255), 2)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 166 |
|
| 167 |
-
# Draw crease line
|
| 168 |
-
cv2.line(frame, (
|
|
|
|
| 169 |
|
| 170 |
if i in detection_frames and trajectory_points.size > 0:
|
| 171 |
idx = detection_frames.index(i) + 1
|
| 172 |
if idx <= len(trajectory_points):
|
| 173 |
-
cv2.polylines(frame, [trajectory_points[:idx]], False, (
|
| 174 |
|
| 175 |
if pitch_point and i == pitch_frame:
|
| 176 |
x, y = pitch_point
|
| 177 |
-
cv2.circle(frame, (int(x), int(y)), 8, (0, 255, 0), -1)
|
| 178 |
-
cv2.putText(frame, "Pitching
|
| 179 |
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
|
| 180 |
|
| 181 |
if impact_point and i == impact_frame:
|
| 182 |
x, y = impact_point
|
| 183 |
-
cv2.circle(frame, (int(x), int(y)), 8, (0, 0, 255), -1)
|
| 184 |
-
cv2.putText(frame, "Impact
|
| 185 |
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
|
| 186 |
|
| 187 |
if impact_point and i == impact_frame and "Out" in decision:
|
| 188 |
-
cv2.putText(frame, "
|
| 189 |
-
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 165, 255), 1)
|
| 190 |
|
| 191 |
for _ in range(SLOW_MOTION_FACTOR):
|
| 192 |
out.write(frame)
|
|
@@ -212,10 +221,10 @@ iface = gr.Interface(
|
|
| 212 |
inputs=gr.Video(label="Upload Video Clip"),
|
| 213 |
outputs=[
|
| 214 |
gr.Textbox(label="DRS Decision and Debug Log"),
|
| 215 |
-
gr.Video(label="Optimized Slow-Motion Replay with Pitching
|
| 216 |
],
|
| 217 |
title="AI-Powered DRS for LBW in Local Cricket",
|
| 218 |
-
description="Upload a video clip of a cricket delivery to get an LBW decision and optimized slow-motion replay showing pitching
|
| 219 |
)
|
| 220 |
|
| 221 |
if __name__ == "__main__":
|
|
|
|
| 4 |
import torch
|
| 5 |
from ultralytics import YOLO
|
| 6 |
import gradio as gr
|
| 7 |
+
from scipy.interpolate import interp1d, UnivariateSpline
|
| 8 |
import uuid
|
| 9 |
import os
|
| 10 |
|
|
|
|
| 16 |
FRAME_RATE = 20 # Input video frame rate
|
| 17 |
SLOW_MOTION_FACTOR = 2 # Reduced for faster output
|
| 18 |
CONF_THRESHOLD = 0.25 # Confidence threshold for detection
|
| 19 |
+
PITCH_ZONE_Y = 0.85 # Adjusted for pitch near stumps
|
| 20 |
+
IMPACT_ZONE_Y = 0.75 # Adjusted for impact near batsman leg
|
| 21 |
+
IMPACT_DELTA_Y = 30 # Reduced for finer impact detection
|
| 22 |
STUMPS_HEIGHT = 0.711 # meters (height of stumps)
|
| 23 |
|
| 24 |
def process_video(video_path):
|
|
|
|
| 61 |
return None, None, None, None, None, None, "Error: Fewer than 2 valid single-ball detections for trajectory"
|
| 62 |
frame_height = frames[0].shape[0]
|
| 63 |
|
| 64 |
+
# Filter to unique positions
|
| 65 |
unique_positions = [ball_positions[0]]
|
| 66 |
for pos in ball_positions[1:]:
|
| 67 |
if abs(pos[0] - unique_positions[-1][0]) > 10 or abs(pos[1] - unique_positions[-1][1]) > 10:
|
|
|
|
| 70 |
y_coords = [pos[1] for pos in unique_positions]
|
| 71 |
times = np.array([i / FRAME_RATE for i in range(len(unique_positions))])
|
| 72 |
|
| 73 |
+
# Smooth coordinates with spline interpolation
|
| 74 |
+
x_smooth = UnivariateSpline(times, x_coords, s=10)
|
| 75 |
+
y_smooth = UnivariateSpline(times, y_coords, s=10)
|
| 76 |
+
|
| 77 |
pitch_idx = 0
|
| 78 |
for i, y in enumerate(y_coords):
|
| 79 |
if y > frame_height * PITCH_ZONE_Y:
|
|
|
|
| 98 |
times = times[:impact_idx + 1]
|
| 99 |
|
| 100 |
try:
|
| 101 |
+
fx = interp1d(times, x_smooth(times), kind='linear', fill_value="extrapolate")
|
| 102 |
+
fy = interp1d(times, y_smooth(times), kind='quadratic', fill_value="extrapolate")
|
| 103 |
except Exception as e:
|
| 104 |
return None, None, None, None, None, None, f"Error in trajectory interpolation: {str(e)}"
|
| 105 |
|
| 106 |
+
vis_trajectory = list(zip(x_smooth(times), y_smooth(times)))
|
| 107 |
t_full = np.linspace(times[0], times[-1] + 0.5, len(times) + 5)
|
| 108 |
x_full = fx(t_full)
|
| 109 |
y_full = fy(t_full)
|
|
|
|
| 122 |
|
| 123 |
frame_height, frame_width = frames[0].shape[:2]
|
| 124 |
stumps_x = frame_width / 2
|
| 125 |
+
stumps_y = frame_height * 0.85 # Adjusted to align with pitch
|
| 126 |
stumps_width_pixels = frame_width * (STUMPS_WIDTH / 3.0)
|
| 127 |
+
batsman_area_y = frame_height * 0.75
|
| 128 |
|
| 129 |
pitch_x, pitch_y = pitch_point
|
| 130 |
impact_x, impact_y = impact_point
|
|
|
|
| 154 |
return None
|
| 155 |
frame_height, frame_width = frames[0].shape[:2]
|
| 156 |
stumps_x = frame_width / 2
|
| 157 |
+
stumps_y = frame_height * 0.85 # Align with pitch
|
| 158 |
stumps_width_pixels = frame_width * (STUMPS_WIDTH / 3.0)
|
| 159 |
stumps_height_pixels = frame_height * (STUMPS_HEIGHT / 3.0)
|
| 160 |
|
|
|
|
| 164 |
trajectory_points = np.array(vis_trajectory, dtype=np.int32).reshape((-1, 1, 2))
|
| 165 |
|
| 166 |
for i, frame in enumerate(frames):
|
| 167 |
+
# Draw stumps outline
|
| 168 |
cv2.line(frame, (int(stumps_x - stumps_width_pixels / 2), int(stumps_y)),
|
| 169 |
(int(stumps_x + stumps_width_pixels / 2), int(stumps_y)), (255, 255, 255), 2)
|
| 170 |
+
cv2.line(frame, (int(stumps_x - stumps_width_pixels / 2), int(stumps_y - stumps_height_pixels)),
|
| 171 |
+
(int(stumps_x - stumps_width_pixels / 2), int(stumps_y)), (255, 255, 255), 2)
|
| 172 |
+
cv2.line(frame, (int(stumps_x + stumps_width_pixels / 2), int(stumps_y - stumps_height_pixels)),
|
| 173 |
+
(int(stumps_x + stumps_width_pixels / 2), int(stumps_y)), (255, 255, 255), 2)
|
| 174 |
|
| 175 |
+
# Draw crease line at stumps
|
| 176 |
+
cv2.line(frame, (int(stumps_x - stumps_width_pixels / 2), int(stumps_y)),
|
| 177 |
+
(int(stumps_x + stumps_width_pixels / 2), int(stumps_y)), (255, 255, 0), 2)
|
| 178 |
|
| 179 |
if i in detection_frames and trajectory_points.size > 0:
|
| 180 |
idx = detection_frames.index(i) + 1
|
| 181 |
if idx <= len(trajectory_points):
|
| 182 |
+
cv2.polylines(frame, [trajectory_points[:idx]], False, (0, 0, 255), 2) # Blue trajectory
|
| 183 |
|
| 184 |
if pitch_point and i == pitch_frame:
|
| 185 |
x, y = pitch_point
|
| 186 |
+
cv2.circle(frame, (int(x), int(y)), 8, (0, 255, 0), -1) # Green for pitching
|
| 187 |
+
cv2.putText(frame, "Pitching", (int(x) + 10, int(y) - 10),
|
| 188 |
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
|
| 189 |
|
| 190 |
if impact_point and i == impact_frame:
|
| 191 |
x, y = impact_point
|
| 192 |
+
cv2.circle(frame, (int(x), int(y)), 8, (0, 0, 255), -1) # Red for impact
|
| 193 |
+
cv2.putText(frame, "Impact", (int(x) + 10, int(y) + 20),
|
| 194 |
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
|
| 195 |
|
| 196 |
if impact_point and i == impact_frame and "Out" in decision:
|
| 197 |
+
cv2.putText(frame, "Wickets", (int(stumps_x) - 50, int(stumps_y) - 20),
|
| 198 |
+
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 165, 255), 1) # Orange for wickets
|
| 199 |
|
| 200 |
for _ in range(SLOW_MOTION_FACTOR):
|
| 201 |
out.write(frame)
|
|
|
|
| 221 |
inputs=gr.Video(label="Upload Video Clip"),
|
| 222 |
outputs=[
|
| 223 |
gr.Textbox(label="DRS Decision and Debug Log"),
|
| 224 |
+
gr.Video(label="Optimized Slow-Motion Replay with Pitching (Green), Impact (Red), Wickets (Orange), Stumps (White), Crease (Yellow)")
|
| 225 |
],
|
| 226 |
title="AI-Powered DRS for LBW in Local Cricket",
|
| 227 |
+
description="Upload a video clip of a cricket delivery to get an LBW decision and optimized slow-motion replay showing pitching (green circle), impact (red circle), wickets (orange text), stumps (white outline), and crease line (yellow line)."
|
| 228 |
)
|
| 229 |
|
| 230 |
if __name__ == "__main__":
|