Printing the Underdetermined: Materializing Multi-solutionness in Figurative Paintings
Yutao Ming, Teng Xu, Youjia Wang, Yunyang Liu, Fengmin Yang
実装難易度
Hard
推論・学習コスト
High
想定用途
生成
概要
Figurative paintings are often approached as if they depict a single recoverable 3D scene: viewers infer depth and occlusion, and reconstruction pipelines attempt to converge to one stable model. We instead foreground multi-solutionness, the non-uniqueness of 3D configurations compatible with a single painted image, and propose a workflow that keeps this non-uniqueness visible and material.…
何が新しいか
Figurative paintings are often approached as if they depict a single recoverable 3D scene: viewers infer depth and occlusion, and reconstruction pipelines attempt to converge to one stable model. We…
何に使えるか
生成
実装情報
- Paper URL
- あり
実装チェックリスト
実装または配布ページ
要確認Paper onlyの可能性があるため再実装前提で確認してください。
一次情報リンク
OKPaper
検証しやすさ
要確認公式実装が見つからないため、論文から再実装する前提です。
計算資源
要確認学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
ライセンス
未取得配布元のLICENSE、モデルカード、Paperの利用条件を確認してください。
商用利用
未取得研究利用限定、データセット由来制限、API規約の有無を確認してください。
自社データで試すなら
製造業・材料開発のExcel/CSVデータに落とし込むための最初の手順です。
- 1まず自社データを、入力条件、目的変数、評価したい指標に分けて整理します。
- 2LightGBMやRandom Forestなどのベースラインを先に作り、この手法と比較します。
- 3評価指標はR2/RMSE、AUC、異常検知の再現率、実験回数削減率など、現場の意思決定に近いものを選びます。
- 4SHAPや特徴量重要度で、効いている因子が物理・化学・工程知識と矛盾しないか確認します。
実装難易度
Hard - 公式実装が見つからないため、論文から再実装する前提です。
必要リソース
- GPU目安: High
- データセット: 論文・リポジトリ側の指定を確認してください。
- 学習要否: 再学習や評価環境の準備が必要になる可能性があります。
- 学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
実務で使う場合の注意点
- ライセンスと商用利用条件は、Paper / GitHub / Hugging Face の配布元で確認してください。
- 精度、再現性、計算コストはデータセットや評価条件に依存します。
- 個人情報や機密データを扱う場合は、入力データの保存先と外部API利用条件を確認してください。
関連記事
Sana — SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformer
SANAは、高解像度画像生成モデルSANAを紹介する本研究であり、低計算コストで優れた高解像度画像を生成できる。
WZPlanner: Safe End-to-End Path Planning for Autonomous Driving in Work Zones
Work zones alter lane geometry through temporary traffic controls and closures that may be absent from on-boar
LynnReal-Omni: Native multi-modal Video Generation for Agentic Visual Workflows
Video diffusion models are stochastic and hard to control: precise content often requires repeated sampling wi
Astronex-World 1.0: Real-Time Interactive World Model Foundation
We present Astronex-World 1.0, an open controllable video world-model foundation. Given a text prompt (text-to