app.py

import streamlit as st
from ultralytics import YOLO
from ultralytics.nn.tasks import SegmentationModel
import torch
import cv2
import numpy as np
import tempfile
import os

# Fix for PyTorch 2.6
torch.serialization.add_safe_globals([SegmentationModel])

# Load YOLOv8 segmentation model on CPU
model = YOLO("yolov8n-seg.pt").to("cpu")

st.title("🎥 Object Segmentation on Uploaded Video")

uploaded_video = st.file_uploader("Upload a video", type=["mp4", "avi", "mov"])

if uploaded_video:
    # Save video temporarily
    tfile = tempfile.NamedTemporaryFile(delete=False)
    tfile.write(uploaded_video.read())
    video_path = tfile.name

    st.video(video_path)
    st.markdown("### ⏳ Processing video... Please wait.")

    cap = cv2.VideoCapture(video_path)
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = int(cap.get(cv2.CAP_PROP_FPS))

    output_path = os.path.join(tempfile.gettempdir(), "output_segmentation.mp4")
    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
    out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))

    frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
    progress_bar = st.progress(0)

    # Generate class color map
    np.random.seed(42)
    colors = {i: tuple(np.random.randint(0, 256, 3).tolist()) for i in range(80)}

    frame_index = 0
    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break

        try:
            results = model.predict(frame, conf=0.3, iou=0.5)
        except Exception as e:
            st.error(f"Prediction failed on frame {frame_index}: {e}")
            break

        if results[0].masks is not None:
            masks = results[0].masks.data.cpu().numpy()
            class_ids = results[0].boxes.cls.cpu().numpy().astype(int)
            boxes = results[0].boxes.xyxy.cpu().numpy()
            names = results[0].names

            for mask, class_id, box in zip(masks, class_ids, boxes):
                label = names[class_id]
                color = colors.get(class_id, (0, 255, 0))

                # Resize and apply mask
                resized_mask = cv2.resize(mask, (frame.shape[1], frame.shape[0]))
                mask_bool = resized_mask > 0.5

                # Overlay
                colored_mask = np.zeros_like(frame, dtype=np.uint8)
                colored_mask[mask_bool] = color
                frame = cv2.addWeighted(frame, 1.0, colored_mask, 0.5, 0)

                # Label
                x1, y1, x2, y2 = box.astype(int)
                font = cv2.FONT_HERSHEY_SIMPLEX
                font_scale = 0.6
                thickness = 1
                label_text = label

                (text_width, text_height), _ = cv2.getTextSize(label_text, font, font_scale, thickness)
                text_x = x1
                text_y = y1 - 10 if y1 - 10 > 10 else y1 + text_height + 10

                # Draw label background
                cv2.rectangle(frame, (text_x - 2, text_y - text_height - 4),
                              (text_x + text_width + 2, text_y + 4), (0, 0, 0), -1)
                cv2.putText(frame, label_text, (text_x, text_y),
                            font, font_scale, (255, 255, 255), thickness=1, lineType=cv2.LINE_AA)

        out.write(frame)
        frame_index += 1
        progress_bar.progress(min(frame_index / frame_count, 1.0))

    cap.release()
    out.release()
    progress_bar.empty()

    st.success("✅ Video processing completed!")

    with open(output_path, "rb") as f:
        st.download_button("📥 Download Segmented Video", f, file_name="segmented_output.mp4", mime="video/mp4")