\n\n
\n\n
\n\n
\n Create sumission distributed by the most frequent position in the mask\n
","metadata":{}},{"cell_type":"code","source":"np.random.seed(73)\n\ndef mask_distribution():\n all_norm_positions = []\n heatmap_size = (100, 100)\n heatmap = np.zeros(heatmap_size, dtype=np.float32)\n \n train_masks_dir = '/kaggle/input/recodai-luc-scientific-image-forgery-detection/train_masks'\n \n if not os.path.exists(train_masks_dir):\n return (0.5, 0.5), None\n \n for mask_file in os.listdir(train_masks_dir):\n if mask_file.endswith('.npy'):\n mask_path = os.path.join(train_masks_dir, mask_file)\n try:\n mask = np.load(mask_path)\n \n if mask.ndim == 3:\n if mask.shape[0] == 1:\n mask = mask[0]\n elif mask.shape[2] == 1:\n mask = mask[:, :, 0]\n else:\n mask = (mask == 1).astype(np.uint8)\n if mask.ndim == 3:\n mask = mask[:, :, 0] if mask.shape[2] == 1 else mask[:, :, 0]\n \n if mask.ndim != 2:\n continue\n \n y_coords, x_coords = np.where(mask > 0)\n \n if len(y_coords) > 0:\n height, width = mask.shape\n \n for y, x in zip(y_coords, x_coords):\n norm_y = y / height\n norm_x = x / width\n \n heatmap_y = int(norm_y * heatmap_size[0])\n heatmap_x = int(norm_x * heatmap_size[1])\n \n heatmap_y = min(heatmap_y, heatmap_size[0] - 1)\n heatmap_x = min(heatmap_x, heatmap_size[1] - 1)\n \n heatmap[heatmap_y, heatmap_x] += 1\n all_norm_positions.append((norm_x, norm_y))\n \n except Exception as e:\n continue\n \n if all_norm_positions:\n max_heatmap_pos = np.unravel_index(np.argmax(heatmap), heatmap.shape)\n max_norm_y = max_heatmap_pos[0] / heatmap_size[0]\n max_norm_x = max_heatmap_pos[1] / heatmap_size[1]\n return (max_norm_x, max_norm_y), heatmap\n \n return (0.5, 0.5), heatmap\n\nhottest_norm_pos, heatmap = mask_distribution()\n\nif heatmap is not None:\n plt.figure(figsize=(10, 8))\n plt.imshow(heatmap, cmap='hot', interpolation='nearest')\n plt.colorbar()\n plt.title('Forgery Location Heatmap')\n plt.xlabel('Normalized X')\n plt.ylabel('Normalized Y')\n plt.show()\n\ntest_images_dir = '/kaggle/input/recodai-luc-scientific-image-forgery-detection/test_images'\nsample_submission = pd.read_csv('/kaggle/input/recodai-luc-scientific-image-forgery-detection/sample_submission.csv')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-10-24T00:17:15.091557Z","iopub.execute_input":"2025-10-24T00:17:15.091887Z","iopub.status.idle":"2025-10-24T00:17:15.132384Z","shell.execute_reply.started":"2025-10-24T00:17:15.091867Z","shell.execute_reply":"2025-10-24T00:17:15.13126Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"submission_data = []\nfor case_id in sample_submission['case_id']:\n img_path = os.path.join(test_images_dir, f\"{case_id}.png\")\n \n with Image.open(img_path) as img:\n width, height = img.size\n \n if np.random.random() < 0.01:\n mask = np.zeros((height, width), dtype=np.uint8)\n \n offset_x = np.random.uniform(-0.3, 0.3) * width\n offset_y = np.random.uniform(-0.3, 0.3) * height\n \n center_x = int(hottest_norm_pos[0] * width + offset_x)\n center_y = int(hottest_norm_pos[1] * height + offset_y)\n \n h = 4\n w = 4\n \n y0 = max(0, center_y - h//2)\n x0 = max(0, center_x - w//2)\n y1 = min(height, y0 + h)\n x1 = min(width, x0 + w)\n \n actual_h = y1 - y0\n actual_w = x1 - x0\n \n if actual_h > 0 and actual_w > 0:\n mask[y0:y1, x0:x1] = 1\n \n RLE_res = rle_encode(mask)\n res = [int(x) for x in RLE_res]\n annotation = json.dumps(res)\n else:\n annotation = 'authentic'\n \n submission_data.append({\n 'case_id': case_id,\n 'annotation': annotation\n })\n\nsubmission = pd.DataFrame(submission_data)\nsubmission.to_csv('submission.csv', index=False)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"
\n\n
\n\n
\n\n
\n If i have mistake write pls in comments\n
","metadata":{}},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}