Delete asmk-mast3r-ps2-gs-kg-02.ipynb

asmk-mast3r-ps2-gs-kg-02.ipynb

@@ -1 +0,0 @@
-{"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.12.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"colab":{"provenance":[],"gpuType":"T4"},"accelerator":"GPU","kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":14554378,"sourceType":"datasetVersion","datasetId":1429416}],"dockerImageVersionId":31236,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# **asmk-mast3r-ps2-gs-kg** \n\n","metadata":{"id":"qDQLX3PArmh8"}},{"cell_type":"markdown","source":"https://www.kaggle.com/code/stpeteishii/dino-mast3r-gs-kg-34","metadata":{}},{"cell_type":"code","source":"!pip uninstall -y numpy scipy trimesh\n!pip install numpy==1.24.3 scipy==1.11.4 trimesh\n!pip install torch torchvision opencv-python pillow tqdm pyaml cython\n!pip install pycolmap\nbreak","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# restart, then run after","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-01-21T07:23:37.733719Z","iopub.execute_input":"2026-01-21T07:23:37.734014Z","iopub.status.idle":"2026-01-21T07:23:37.738453Z","shell.execute_reply.started":"2026-01-21T07:23:37.733988Z","shell.execute_reply":"2026-01-21T07:23:37.737659Z"}},"outputs":[],"execution_count":1},{"cell_type":"code","source":"\"\"\"\nComplete MASt3R to Gaussian Splatting Pipeline\nProcess2 Only - Memory Optimized Version\n\"\"\"\n\nimport os\nimport sys\nimport gc\nimport torch\nimport numpy as np\nfrom pathlib import Path\nfrom tqdm import tqdm\nimport torch.nn.functional as F\n\n# ======================================================================\n# MEMORY MANAGEMENT\n# ======================================================================\n\nos.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'expandable_segments:True'\n\ndef clear_memory():\n    \"\"\"メモリクリア関数\"\"\"\n    gc.collect()\n    if torch.cuda.is_available():\n        torch.cuda.empty_cache()\n        torch.cuda.synchronize()\n\n\n# ======================================================================\n# CONFIGURATION\n# ======================================================================\n\nclass Config:\n    DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n    MAST3R_WEIGHTS = \"naver/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric\"\n    RETRIEVAL_TOPK = 10\n    IMAGE_SIZE = 224  # メモリ節約のため224に設定\n\n\n# ======================================================================\n# IMAGE LOADING\n# ======================================================================\n\ndef load_images_from_directory(image_dir, max_images=200):\n    \"\"\"ディレクトリから画像をロード\"\"\"\n    print(f\"\\nLoading images from: {image_dir}\")\n    \n    valid_extensions = {'.jpg', '.jpeg', '.png', '.bmp'}\n    image_paths = []\n    \n    for ext in valid_extensions:\n        image_paths.extend(sorted(Path(image_dir).glob(f'*{ext}')))\n        image_paths.extend(sorted(Path(image_dir).glob(f'*{ext.upper()}')))\n    \n    image_paths = sorted(set(str(p) for p in image_paths))\n    \n    if len(image_paths) > max_images:\n        print(f\"⚠️  Limiting from {len(image_paths)} to {max_images} images\")\n        image_paths = image_paths[:max_images]\n    \n    print(f\"✓ Found {len(image_paths)} images\")\n    return image_paths\n\n\n# ======================================================================\n# MAST3R MODEL\n# ======================================================================\n\ndef load_mast3r_model(device):\n    \"\"\"MASt3Rモデルをロード\"\"\"\n    print(\"\\n=== Loading MASt3R Model ===\")\n    \n    # mast3rのパスを追加\n    if '/kaggle/working/mast3r' not in sys.path:\n        sys.path.insert(0, '/kaggle/working/mast3r')\n    if '/kaggle/working/mast3r/dust3r' not in sys.path:\n        sys.path.insert(0, '/kaggle/working/mast3r/dust3r')\n    \n    from dust3r.model import AsymmetricCroCo3DStereo\n    \n    model = AsymmetricCroCo3DStereo.from_pretrained(Config.MAST3R_WEIGHTS).to(device)\n    model.eval()\n    \n    print(f\"✓ Model loaded on {device}\")\n    return model\n\n\n# ======================================================================\n# FEATURE EXTRACTION & PAIR SELECTION\n# ======================================================================\n\ndef load_asmk_retrieval_model(device):\n    \"\"\"ASMKリトリーバルモデルをロード\"\"\"\n    print(\"\\n=== Loading ASMK Retrieval Model ===\")\n    \n    # mast3rとasmkのパスを追加\n    if '/kaggle/working/mast3r' not in sys.path:\n        sys.path.insert(0, '/kaggle/working/mast3r')\n    if '/kaggle/working/mast3r/dust3r' not in sys.path:\n        sys.path.insert(0, '/kaggle/working/mast3r/dust3r')\n    if '/kaggle/working/asmk' not in sys.path:\n        sys.path.insert(0, '/kaggle/working/asmk')\n    \n    from dust3r.model import AsymmetricCroCo3DStereo\n    \n    model = AsymmetricCroCo3DStereo.from_pretrained(Config.MAST3R_WEIGHTS).to(device)\n    model.eval()\n    \n    # Codebookの初期化（簡易版）\n    codebook = np.random.randn(1024, 24).astype(np.float32)\n    \n    print(\"✓ ASMK model loaded\")\n    return model, codebook\n\n\ndef extract_mast3r_features(model, image_paths, device, batch_size=1):\n    \"\"\"MASt3Rモデルを使用して特徴量を抽出\"\"\"\n    print(\"\\n=== Extracting MASt3R Features ===\")\n    from dust3r.utils.image import load_images\n    from dust3r.inference import inference\n    \n    all_features = []\n    \n    for i in tqdm(range(0, len(image_paths), batch_size), desc=\"Features\"):\n        batch_paths = image_paths[i:i+batch_size]\n        images = load_images(batch_paths, size=Config.IMAGE_SIZE)\n        \n        with torch.no_grad():\n            output = inference([tuple(images)], model, device, batch_size=len(images))\n        \n        for idx in range(len(batch_paths)):\n            if 'desc' in output[0][idx]:\n                desc = output[0][idx]['desc']\n            elif 'pts3d' in output[0][idx]:\n                desc = output[0][idx]['pts3d']\n            else:\n                desc = list(output[0][idx].values())[0]\n            \n            if desc.dim() == 4:\n                desc = desc.squeeze(0)\n            \n            all_features.append(desc)\n        \n        del output, images\n        torch.cuda.empty_cache()\n    \n    print(f\"✓ Extracted features for {len(all_features)} images\")\n    if all_features:\n        print(f\"   Feature shape: {all_features[0].shape}\")\n    \n    return all_features\n\n\ndef compute_asmk_similarity(features, codebook):\n    \"\"\"ASMKを使用して類似度行列を計算\"\"\"\n    print(\"\\n=== Computing ASMK Similarity ===\")\n    \n    n_images = len(features)\n    similarity_matrix = np.zeros((n_images, n_images), dtype=np.float32)\n    \n    # 各特徴量をグローバル記述子に変換\n    global_features = []\n    \n    for feat in features:\n        if isinstance(feat, torch.Tensor):\n            feat = feat.cpu().numpy()\n        \n        h, w, c = feat.shape\n        feat_flat = feat.reshape(-1, c)\n        global_desc = np.mean(feat_flat, axis=0)\n        global_features.append(global_desc)\n    \n    global_features = np.stack(global_features)\n    \n    if codebook is not None and len(codebook) > 0:\n        try:\n            print(f\"Using ASMK with codebook size: {len(codebook)}\")\n            \n            for i in range(n_images):\n                feat_i = features[i]\n                if isinstance(feat_i, torch.Tensor):\n                    feat_i = feat_i.cpu().numpy()\n                feat_i = feat_i.reshape(-1, feat_i.shape[-1])\n                \n                for j in range(i+1, n_images):\n                    feat_j = features[j]\n                    if isinstance(feat_j, torch.Tensor):\n                        feat_j = feat_j.cpu().numpy()\n                    feat_j = feat_j.reshape(-1, feat_j.shape[-1])\n                    \n                    dist_i = np.linalg.norm(feat_i[:, None, :] - codebook[None, :, :], axis=2)\n                    dist_j = np.linalg.norm(feat_j[:, None, :] - codebook[None, :, :], axis=2)\n                    \n                    assign_i = np.argmin(dist_i, axis=1)\n                    assign_j = np.argmin(dist_j, axis=1)\n                    \n                    common = len(set(assign_i) & set(assign_j))\n                    sim = common / max(len(set(assign_i)), len(set(assign_j)))\n                    \n                    similarity_matrix[i, j] = sim\n                    similarity_matrix[j, i] = sim\n                \n                if (i + 1) % 10 == 0:\n                    print(f\"Processed {i+1}/{n_images} images\")\n            \n            print(\"✓ ASMK similarity computation completed\")\n            \n        except Exception as e:\n            print(f\"⚠️  ASMK failed: {e}, using cosine similarity\")\n            global_features_norm = global_features / (np.linalg.norm(global_features, axis=1, keepdims=True) + 1e-8)\n            similarity_matrix = global_features_norm @ global_features_norm.T\n    \n    else:\n        print(\"No codebook provided, using cosine similarity\")\n        global_features_norm = global_features / (np.linalg.norm(global_features, axis=1, keepdims=True) + 1e-8)\n        similarity_matrix = global_features_norm @ global_features_norm.T\n    \n    np.fill_diagonal(similarity_matrix, -1)\n    \n    print(f\"Similarity matrix shape: {similarity_matrix.shape}\")\n    print(f\"Similarity range: [{similarity_matrix.min():.3f}, {similarity_matrix.max():.3f}]\")\n    \n    return similarity_matrix\n\n\ndef build_pairs_from_similarity(similarity_matrix, top_k=10):\n    \"\"\"類似度行列からペアを構築\"\"\"\n    n_images = similarity_matrix.shape[0]\n    pairs = []\n    \n    for i in range(n_images):\n        similarities = similarity_matrix[i]\n        top_indices = np.argsort(similarities)[::-1][:top_k]\n        \n        for j in top_indices:\n            if j > i:\n                pairs.append((i, j))\n    \n    pairs = list(set(pairs))\n    print(f\"✓ Built {len(pairs)} unique pairs\")\n    \n    return pairs\n\n\ndef get_image_pairs_asmk(image_paths, max_pairs=100):\n    \"\"\"ASMKを使用して画像ペアを取得\"\"\"\n    print(\"\\n=== Getting Image Pairs with ASMK ===\")\n    \n    device = Config.DEVICE\n    model, codebook = load_asmk_retrieval_model(device)\n    features = extract_mast3r_features(model, image_paths, device)\n    similarity_matrix = compute_asmk_similarity(features, codebook)\n    pairs = build_pairs_from_similarity(similarity_matrix, Config.RETRIEVAL_TOPK)\n    \n    # モデルを解放\n    del model\n    clear_memory()\n    \n    if len(pairs) > max_pairs:\n        pairs = pairs[:max_pairs]\n        print(f\"Limited to {max_pairs} pairs\")\n    \n    return pairs\n\n\n# ======================================================================\n# MAST3R RECONSTRUCTION\n# ======================================================================\n\ndef run_mast3r_pairs(model, image_paths, pairs, device, batch_size=1):\n    \"\"\"MASt3Rでペア画像を処理（メモリ最適化版）\"\"\"\n    print(\"\\n=== Running MASt3R Reconstruction ===\")\n    from dust3r.inference import inference\n    from dust3r.cloud_opt import global_aligner, GlobalAlignerMode\n    from dust3r.utils.image import load_images\n    \n    # ペアを制限（メモリ節約）\n    max_pairs_for_memory = 50\n    if len(pairs) > max_pairs_for_memory:\n        print(f\"⚠️  Limiting pairs from {len(pairs)} to {max_pairs_for_memory} for memory\")\n        pairs = pairs[:max_pairs_for_memory]\n    \n    # ペアから画像インデックスを取得\n    pair_indices = []\n    for i, j in pairs:\n        pair_indices.extend([i, j])\n    unique_indices = sorted(set(pair_indices))\n    \n    selected_paths = [image_paths[i] for i in unique_indices]\n    print(f\"Selected {len(selected_paths)} unique images from {len(pairs)} pairs\")\n    \n    # 画像をロード\n    images = load_images(selected_paths, size=Config.IMAGE_SIZE)\n    \n    clear_memory()\n    \n    # バッチサイズを動的に調整\n    available_memory = torch.cuda.get_device_properties(device).total_memory\n    used_memory = torch.cuda.memory_allocated(device)\n    free_memory = available_memory - used_memory\n    \n    if free_memory < 2e9:\n        batch_size = 1\n        print(f\"⚠️  Low memory, using batch_size=1\")\n    \n    # 推論を小さなバッチで実行\n    all_outputs = []\n    for i in range(0, len(images), batch_size):\n        batch_images = images[i:i+batch_size]\n        print(f\"Processing batch {i//batch_size + 1}/{(len(images)-1)//batch_size + 1}\")\n        \n        with torch.no_grad():\n            output = inference([tuple(batch_images)], model, device, batch_size=len(batch_images))\n        \n        all_outputs.extend(output[0])\n        clear_memory()\n    \n    print(f\"✓ Processed {len(all_outputs)} image predictions\")\n    \n    # Global alignmentの準備\n    scene = global_aligner(\n        dust3r_output=all_outputs,\n        device=device,\n        mode=GlobalAlignerMode.PointCloudOptimizer,\n        verbose=True\n    )\n    \n    clear_memory()\n    \n    # Global alignment\n    print(\"Running global alignment...\")\n    try:\n        loss = scene.compute_global_alignment(\n            init=\"mst\", \n            niter=50,\n            schedule='cosine', \n            lr=0.01\n        )\n        print(f\"✓ Alignment complete (loss: {loss:.6f})\")\n    except RuntimeError as e:\n        if \"out of memory\" in str(e).lower():\n            print(\"⚠️  OOM during alignment, trying with fewer iterations...\")\n            clear_memory()\n            \n            loss = scene.compute_global_alignment(\n                init=\"mst\", \n                niter=20,\n                schedule='cosine', \n                lr=0.01\n            )\n            print(f\"✓ Alignment complete with reduced iterations (loss: {loss:.6f})\")\n        else:\n            raise\n    \n    clear_memory()\n    \n    return scene, images\n\n\n# ======================================================================\n# CAMERA PARAMETER EXTRACTION (PROCESS2)\n# ======================================================================\n\ndef extract_camera_params_process2(scene, image_paths, conf_threshold=1.5):\n    \"\"\"Process2: sceneから直接カメラパラメータと3D点を抽出\"\"\"\n    print(\"\\n=== Extracting Camera Parameters (Process2) ===\")\n    \n    cameras_dict = {}\n    all_pts3d = []\n    all_confidence = []\n    \n    n_images = len(image_paths)\n    \n    for idx in range(n_images):\n        img_name = os.path.basename(image_paths[idx])\n        \n        try:\n            # カメラパラメータを取得\n            if hasattr(scene, 'im_poses'):\n                pose = scene.im_poses[idx]\n            elif hasattr(scene, 'get_im_poses'):\n                pose = scene.get_im_poses()[idx]\n            else:\n                pose = torch.eye(4)\n            \n            if hasattr(scene, 'im_focals'):\n                focal = scene.im_focals[idx]\n            elif hasattr(scene, 'get_focals'):\n                focal = scene.get_focals()[idx]\n            else:\n                focal = 1000.0\n            \n            if hasattr(scene, 'im_pp'):\n                pp = scene.im_pp[idx]\n            elif hasattr(scene, 'get_principal_points'):\n                pp = scene.get_principal_points()[idx]\n            else:\n                pp = torch.tensor([112.0, 112.0])\n            \n            # テンソルをnumpyに変換\n            if isinstance(pose, torch.Tensor):\n                pose = pose.detach().cpu().numpy()\n            if isinstance(focal, torch.Tensor):\n                focal = focal.detach().cpu().item()\n            if isinstance(pp, torch.Tensor):\n                pp = pp.detach().cpu().numpy()\n            \n            # カメラパラメータを保存\n            cameras_dict[img_name] = {\n                'focal': focal,\n                'pp': pp,\n                'pose': pose\n            }\n            \n            # 3D点を取得\n            if hasattr(scene, 'im_pts3d'):\n                pts3d_img = scene.im_pts3d[idx]\n            elif hasattr(scene, 'get_pts3d'):\n                pts3d_img = scene.get_pts3d()[idx]\n            else:\n                pts3d_img = None\n            \n            # Confidenceを取得\n            if hasattr(scene, 'im_conf'):\n                conf_img = scene.im_conf[idx]\n            elif hasattr(scene, 'get_conf'):\n                conf_img = scene.get_conf()[idx]\n            else:\n                conf_img = None\n            \n            # 3D点とconfidenceを処理\n            if pts3d_img is not None:\n                if isinstance(pts3d_img, torch.Tensor):\n                    pts3d_img = pts3d_img.detach().cpu().numpy()\n                \n                if pts3d_img.ndim == 3:\n                    pts3d_flat = pts3d_img.reshape(-1, 3)\n                else:\n                    pts3d_flat = pts3d_img\n                \n                all_pts3d.append(pts3d_flat)\n                \n                # confidenceを処理\n                if conf_img is not None:\n                    if isinstance(conf_img, list):\n                        conf_img = np.array(conf_img)\n                    elif isinstance(conf_img, torch.Tensor):\n                        conf_img = conf_img.detach().cpu().numpy()\n                    \n                    if conf_img.ndim > 1:\n                        conf_flat = conf_img.reshape(-1)\n                    else:\n                        conf_flat = conf_img\n                    \n                    if len(conf_flat) != len(pts3d_flat):\n                        conf_flat = np.ones(len(pts3d_flat))\n                    \n                    all_confidence.append(conf_flat)\n                else:\n                    all_confidence.append(np.ones(len(pts3d_flat)))\n        \n        except Exception as e:\n            print(f\"⚠️  Error processing image {idx} ({img_name}): {e}\")\n            cameras_dict[img_name] = {\n                'focal': 1000.0,\n                'pp': np.array([112.0, 112.0]),\n                'pose': np.eye(4)\n            }\n            continue\n    \n    # 全3D点を結合\n    if all_pts3d:\n        pts3d = np.vstack(all_pts3d)\n        confidence = np.concatenate(all_confidence)\n    else:\n        pts3d = np.zeros((0, 3))\n        confidence = np.zeros(0)\n    \n    print(f\"✓ Extracted camera parameters for {len(cameras_dict)} images\")\n    print(f\"✓ Total 3D points: {len(pts3d)}\")\n    print(f\"✓ Confidence shape: {confidence.shape}\")\n    \n    # Confidenceでフィルタリング\n    if len(confidence) > 0:\n        valid_mask = confidence > conf_threshold\n        pts3d = pts3d[valid_mask]\n        confidence = confidence[valid_mask]\n        print(f\"✓ After confidence filtering (>{conf_threshold}): {len(pts3d)} points\")\n    \n    return cameras_dict, pts3d, confidence\n\n\n# ======================================================================\n# COLMAP EXPORT\n# ======================================================================\n\ndef write_colmap_sparse(cameras_dict, pts3d, confidence, image_paths, output_dir):\n    \"\"\"COLMAPフォーマットでスパース再構成を書き出し\"\"\"\n    print(\"\\n=== Writing COLMAP Sparse Reconstruction ===\")\n    \n    import pycolmap\n    \n    os.makedirs(output_dir, exist_ok=True)\n    \n    # Reconstruction オブジェクトを作成\n    reconstruction = pycolmap.Reconstruction()\n    \n    # カメラを追加\n    camera_id = reconstruction.add_camera(\n        pycolmap.Camera(\n            model=\"SIMPLE_PINHOLE\",\n            width=Config.IMAGE_SIZE,\n            height=Config.IMAGE_SIZE,\n            params=[1000.0, Config.IMAGE_SIZE/2, Config.IMAGE_SIZE/2]\n        )\n    )\n    \n    # 画像を追加\n    for img_idx, img_path in enumerate(image_paths):\n        img_name = os.path.basename(img_path)\n        \n        if img_name in cameras_dict:\n            cam_params = cameras_dict[img_name]\n            pose = cam_params['pose']\n            \n            # Rotation and translation\n            R = pose[:3, :3]\n            t = pose[:3, 3]\n            \n            qvec = pycolmap.rotmat_to_qvec(R)\n            tvec = t\n            \n            reconstruction.add_image(\n                pycolmap.Image(\n                    id=img_idx + 1,\n                    name=img_name,\n                    camera_id=camera_id,\n                    qvec=qvec,\n                    tvec=tvec\n                )\n            )\n    \n    # 3D点を追加\n    for i, pt in enumerate(pts3d):\n        if i >= 100000:  # 点数制限\n            break\n        \n        reconstruction.add_point3D(\n            pycolmap.Point3D(\n                xyz=pt,\n                color=np.array([128, 128, 128], dtype=np.uint8),\n                error=1.0 - confidence[i] if i < len(confidence) else 1.0\n            )\n        )\n    \n    # 保存\n    reconstruction.write(output_dir)\n    \n    print(f\"✓ Wrote COLMAP reconstruction to: {output_dir}\")\n    print(f\"  - Cameras: {len(reconstruction.cameras)}\")\n    print(f\"  - Images: {len(reconstruction.images)}\")\n    print(f\"  - Points: {len(reconstruction.points3D)}\")\n\n\n# ======================================================================\n# GAUSSIAN SPLATTING\n# ======================================================================\n\ndef run_gaussian_splatting(output_dir, iterations=30000):\n    \"\"\"Gaussian Splattingを実行\"\"\"\n    print(\"\\n=== Running Gaussian Splatting ===\")\n    \n    gs_source = output_dir\n    gs_model = os.path.join(output_dir, \"output\")\n    \n    cmd = f\"\"\"\n    python /kaggle/working/gaussian-splatting/train.py \\\n        -s {gs_source} \\\n        -m {gs_model} \\\n        --iterations {iterations} \\\n        --eval\n    \"\"\"\n    \n    print(f\"Command: {cmd}\")\n    os.system(cmd)\n    \n    print(f\"✓ Gaussian Splatting complete\")\n    print(f\"  Output: {gs_model}\")\n    \n    return gs_model\n\n\n# ======================================================================\n# MAIN PIPELINE\n# ======================================================================\n\ndef main_pipeline(image_dir, output_dir, square_size=256, iterations=30000, \n                  max_images=200, max_pairs=100, max_points=500000, \n                  conf_threshold=1.5):\n    \"\"\"メインパイプライン（Process2のみ、メモリ最適化版）\"\"\"\n    \n    print(\"=\"*70)\n    print(\"STEP 1: Loading and Preparing Images\")\n    print(\"=\"*70)\n    \n    image_paths = load_images_from_directory(image_dir, max_images=max_images)\n    print(f\"Loaded {len(image_paths)} images\")\n    \n    clear_memory()\n    \n    print(\"\\n\" + \"=\"*70)\n    print(\"STEP 2: Image Pair Selection\")\n    print(\"=\"*70)\n    \n    max_pairs = min(max_pairs, 50)\n    pairs = get_image_pairs_asmk(image_paths, max_pairs=max_pairs)\n    print(f\"Selected {len(pairs)} image pairs\")\n    \n    clear_memory()\n    \n    print(\"\\n\" + \"=\"*70)\n    print(\"STEP 3: MASt3R 3D Reconstruction\")\n    print(\"=\"*70)\n    \n    device = Config.DEVICE\n    model = load_mast3r_model(device)\n    \n    scene, mast3r_images = run_mast3r_pairs(model, image_paths, pairs, device)\n    \n    # モデルを解放\n    del model\n    clear_memory()\n    \n    print(\"\\n\" + \"=\"*70)\n    print(\"STEP 4: Converting to COLMAP (Process2 Method)\")\n    print(\"=\"*70)\n    \n    cameras_dict, pts3d, confidence = extract_camera_params_process2(\n        scene, image_paths, conf_threshold=conf_threshold\n    )\n    \n    # sceneを解放\n    del scene\n    clear_memory()\n    \n    # 点数を制限\n    if len(pts3d) > max_points:\n        print(f\"⚠️  Limiting points from {len(pts3d)} to {max_points}\")\n        indices = np.random.choice(len(pts3d), max_points, replace=False)\n        pts3d = pts3d[indices]\n        confidence = confidence[indices]\n    \n    print(f\"Final point count: {len(pts3d)}\")\n    \n    # COLMAP変換\n    colmap_dir = os.path.join(output_dir, \"sparse/0\")\n    os.makedirs(colmap_dir, exist_ok=True)\n    \n    write_colmap_sparse(cameras_dict, pts3d, confidence, image_paths, colmap_dir)\n    \n    clear_memory()\n    \n    print(\"\\n\" + \"=\"*70)\n    print(\"STEP 5: Running Gaussian Splatting\")\n    print(\"=\"*70)\n    \n    gs_output = run_gaussian_splatting(\n        output_dir=output_dir,\n        iterations=iterations\n    )\n    \n    return gs_output\n\n\n# ======================================================================\n# USAGE EXAMPLE\n# ======================================================================\n\nif __name__ == \"__main__\":\n    IMAGE_DIR = \"/kaggle/input/two-dogs/fountain80/fountain80\"\n    OUTPUT_DIR = \"/kaggle/working/output\"\n    \n    # メモリ制約を考慮したパラメータ\n    gs_output = main_pipeline(\n        image_dir=IMAGE_DIR,\n        output_dir=OUTPUT_DIR,\n        square_size=256,\n        iterations=30000,\n        max_images=100,      # 画像数を制限\n        max_pairs=50,        # ペア数を制限\n        max_points=300000,   # 点数を制限\n        conf_threshold=1.5\n    )\n    \n    print(\"\\n\" + \"=\"*70)\n    print(\"PIPELINE COMPLETE\")\n    print(\"=\"*70)\n    print(f\"Output directory: {gs_output}\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}