JP4733332B2 - Material identification method - Google Patents

Description

【００７０】
表１中の１６進数の値は変更してもよい。これらの値は一例である。また、第１〜第１２バイトは、表中に示すもの以外の指定であってよい。表１に示す例では、第４バイトは、第５〜第１２バイトがＳＭＰＴＥに合致するデータフォーマットに関することを示す。第５バイトは、第６〜第１０バイトが「ディクショナリ」データに関することを示す。第６バイトは、このようなデータが第７〜第１０バイトにより定義される「メタデータ」であることを示す。第７バイトは、第９及び第１０バイトにより定義されるメタデータを含むディクショナリの一部を示す。第８バイトは、ディクショナリのバーションを示す。第９バイトはデータのクラスを示し、第１０バイトはそのクラスにおける特定アイテムを示す。
【００７１】
本実施例では、第１〜第１０バイトが予め割り当てられた固定値を有している。第１１バイトは変数である。図７及び上記の表１を参照して、ＵＭＩＤのラベルの第１〜第１０バイトは固定されていることがわかる。従って、これらのバイトは、第１〜第１０バイトを表す１バイトの「タイプ」コードＴにより置き換えることができる。タイプコードＴの後には長さコードＬが来る。これに続いて、一方が表１の第１１バイトであるとともに他方が表１の第１２バイトである２バイトと、インスタンス番号（３バイト）と、マテリアル番号（１６バイト）がある。また、任意であるが、マテリアル番号の後に拡張ＵＭＩＤのシグネチャメタデータ及び／又は他のメタデータを配置してもよい。
【００７２】
ＵＭＩＤタイプ（第１１バイト）は、以下ように４つの異なるデータタイプをそれぞれ識別するための４つの値を有している。
【００７３】
「０１ｈ」＝ピクチャマテリアルのＵＭＩＤ
「０２ｈ」＝オーディオマテリアルのＵＭＩＤ
「０３ｈ」＝データマテリアルのＵＭＩＤ
「０４ｈ」＝グループマテリアル（例えば、関連する内容の組み合わせ）のＵＭＩＤ
１２バイトのラベルの最後の（第１２）バイトは、マテリアル及びインスタンス番号を作成する方法を識別する。このバイトはトップニブルとボトムニブルに分割され、トップニブルがマテリアル番号作成方法を定義するとともに、ボトムニブルがインスタンス番号作成方法を定義する。
【００７４】
「長さ」
長さは、ベーシックＵＭＩＤについての値「１３ｈ」と拡張ＵＭＩＤについての値「３３ｈ」を有する１バイトの数字である。
【００７５】
「インスタンス番号」
インスタンス番号は、規格で定められた幾つかの手段のうちのいずれにより作成されるユニークな３バイトの番号である。インスタンス番号により、あるクリップの特定の「インスタンス」と外部の関連するメタデータとのリンクが得られる。このインスタンス番号がないと、すべてのマテリアルが、マテリアルのいずれのインスタンスとも、また、それに関連するメタデータとも結びついてしまう。
【００７６】
新たなクリップの作成には、ゼロのインスタンス番号とともに新たなマテリアル番号の作成が必要である。従って、非ゼロのインスタンス番号は、関連するクリップがソースマテリアルではないことを示す。インスタンス番号は主として、あるクリップの特定のインスタンスに関する関連メタデータを識別するのに使用される。
【００７７】
「マテリアル番号」
１６バイトのマテリアル番号は、規格で定められた幾つかの手段のうちのいずれかにより作成されうる非ゼロ番号である。この番号は、６バイトの登録ポートＩＤ番号、時間、乱数発生器によって変わる。
【００７８】
「シグネチャメタデータ」
シグネチャメタデータから成分はいずれも、有意義な値を入れることができない場合には、ヌルフィルを行うことができる。ヌルフィルされた成分はいずれも、後段のデコーダに対してその成分が有効でないことを明確に示すため、完全にヌルフィルされる。
【００７９】
「日時フォーマット」
日時フォーマットは８バイトであり、その最初の４バイトは時間成分に基づくＵＴＣ（ユニバーサルタイムコード）である。時間は、内容のタイプによって、ＡＥＳ３の３２ビットオーディオサンプルクロック又はＳＭＰＴＥ１２Ｍのいずれかにより定められる。
【００８０】
後の４バイトは、ＳＭＰＴＥ３０９Ｍで定義するようなモディファイドジュリアンデータ（ＭＪＤ）に基づくデータを定める。これは、１８５８年１１月１７日午前０時から９９９，９９９日をカウントし、西暦４５９７年までの日付を可能とする。
【００８１】
「空間座標フォーマット」
空間座標値は、以下のように定義される３成分からなる。
・高度：９９，９９９，９９９メートルまでを特定する８個の１０進数。
・経度：東経／西経１８０．０００００度（少数第５位まで有効）を特定する８個の１０進数。
・緯度：北緯／南緯９０．０００００度（少数第５位まで有効）を特定する８個の１０進数。
【００８２】
高度の値は、地球の中心からの値をメートルで表したものなので、海水面より低い高度も可能とする。
【００８３】
なお、空間座標はほとんどのクリップについてスタティックであるが、すべての場合にあてはまるわけではない。乗り物に取り付けたカメラ等の移動ソースから取り込まれたマテリアルの場合、空間座標値が変化することもある。
【００８４】
「国コード」
国コードは、ＩＳＯ３１６６で定めたセットに応じた短縮形の４バイトの英数字ストリングである。登録されていない国は、ＳＭＰＴＥレジストレーションオーソリティから登録英数字ストリングを得ることができる。
【００８５】
「団体コード」
団体コードは、ＳＭＰＴＥに登録された短縮形の４バイトの英数字ストリングである。団体コードは、それらの登録国コードに関してのみ意味を持つので、団体コードは異なる国においても同じ値を有することができる。
【００８６】
「ユーザコード」
ユーザコードは、各団体によりローカル的に割り当てられるが、グローバル的に登録されてはいない４バイトの英数字ストリングである。ユーザコードは、それらの登録団体コード及び国コードに関してのみ定められるので、ユーザコードは異なる団体や国においても同じ値を有することができる。
【００８７】
「フリーランスオペレータ」
フリーランスオペレータは、国コードとして自分の居住国を使用し、例えばＳＭＰＴＥに登録することができる８バイトのコードに結合された団体コード及びユーザコードを使用する。これらのフリーランスコードは、「〜」の記号（ＩＳＯ８８５９文字番号７Ｅｈ）で始まり、その後に登録された７桁の英数字ストリングを有する。
【００８８】
なお、上述の説明から、ＵＭＩＤは、ビデオマテリアルだけでなく、オーディオマテリアル、データマテリアル、マテリアルグループも識別するのに用いることができることがわかる。
【００８９】
メタデータ−図８
以下、プログラムの作成中に生成することが可能なメタデータの種類と、そのメタデータを構成するのに可能な構成アプローチについて、例を用いて説明する。ウォーターマークとしてビデオに埋め込まれたＵＭＩＤは、データベースにおいて関連するメタデータにビデオをリンクするアイデンティファイアとして機能する。
【００９０】
図８は、メタデータを構成するための構造例を示す。メタデータをそれぞれ有する多数のフィールドからなる多数のテーブルが設けられている。これらのテーブルは、各テーブル内の共通のフィールドにより互いに関連付けられることにより、リレーショナル構造を得ることができる。また、この構造は、テーブルが表すオブジェクトの多数のインスタンスを表すために、同じテーブルの多数のインスタンスを有してもよい。フィールドは所定の方法でフォーマットされる。フィールドのサイズも所定のものとすることができる。サイズの例としては、２バイトを表す「Ｉｎｔ」、４バイトを表す「ＬｏｎｇＩｎｔ」、８バイトを表す「Ｄｏｕｂｌｅ」がある。また、フィールドのサイズを、例えば８、１０、１６、３２、１２８、２５５文字等、フィールド内に保持する文字数を参照して定めることもできる。
【００９１】
この構造について詳細に説明すると、プログラムテーブルがある。プログラムテーブルは、プログラムＩＤ（ＰＩＤ）、タイトル、ワーキングタイトル、ジャンルＩＤ、一覧、アスペクト比、ディレクターＩＤ、ピクチャスタンプを含む多数のフィールドを有している。プログラムテーブルに関連して、ジャンルテーブル、キーワードテーブル、スクリプトテーブル、ピープルテーブル、スケジュールテーブル、及び、複数のメディアオブジェクトテーブルがある。
【００９２】
ジャンルテーブルは、プログラムテーブルのジャンルＩＤフィールドと関連付けられたジャンルＩＤと、ジャンルディスクリプションとを含む多数のフィールドを有している。
【００９３】
キーワードテーブルは、プログラムテーブルのプログラムＩＤフィールドと関連付けられたプログラムＩＤと、キーワードＩＤと、キーワードとを含む多数のフィールドを有している。
【００９４】
スクリプトテーブルは、スクリプトＩＤ、スクリプト名、スクリプトタイプ、ドキュメントフォーマット、パス、作成日、原作者、バージョン、最終変更日、変更者、プログラムＩＤに関連付けられたＰＩＤ、ノートと含む多数のフィールドを有している。ピープルテーブルは、イメージを含む多数のフィールドを有している。
【００９５】
ピープルテーブルは、多数の個別テーブルと多数のグループテーブルに関連付けられている。各個別テーブルは、イメージを含む多数のフィールドを有している。各グループテーブルは、イメージを含む多数のフィールドを有している。各個別テーブルは、プロダクションスタッフテーブル又はキャストテーブルのいずれかと関連付けられている。
【００９６】
プロダクションスタッフテーブルは、プロダクションスタッフＩＤ、姓、ファーストネーム、契約ＩＤ、エージェント、エージェンシーＩＤ、Ｅメール、住所、電話番号、役割ＩＤ、ノート、アレルギー、誕生日、国民保険番号、バンクＩＤ、ピクチャスタンプを含む多数のフィールドを有している。
【００９７】
キャストテーブルは、キャストＩＤ、姓、ファーストネーム、役名、契約ＩＤ、エージェント、エージェンシーＩＤ、エクイティ番号、Ｅメール、住所、電話番号、誕生日、バンクＩＤ、ピクチャスタンプを含む多数のフィールドを有している。プロダクションスタッフテーブルとキャストテーブルに関連して、バンクディテールテーブルとエージェンシーテーブルがある。
【００９８】
バンクディテールテーブルは、プロダクションスタッフテーブルのバンクＩＤフィールドとキャストテーブルのバンクＩＤフィールドに関連付けられたバンクＩＤ、ソートコード、口座番号、口座名を含む多数のフィールドを有している。
【００９９】
エージェンシーテーブルは、プロダクションスタッフテーブルのエージェンシーＩＤフィールドとキャストテーブルのエージェンシーＩＤフィールドに関連付けられたエージェンシーＩＤ、名称、所在地、電話番号、ウェブサイト及びＥメール、ピクチャスタンプを含む多数のフィールドを有している。また、プロダクションスタッフテーブルに関連して役割テーブルがある。
【０１００】
役割テーブルは、プロダクションスタッフテーブルの役割ＩＤフィールドに関連付けられた役割ＩＤ、職務、ノート、ピクチャスタンプを含む多数のフィールドを有している。各グループテーブルは、団体テーブルと関連付けられている。
【０１０１】
団体テーブルは、団体ＩＤ、名称、種類、所在地、契約ＩＤ、連絡先名、連絡先電話番号、ウェブサイト、ピクチャスタンプを含む多数のフィールドを有している。
【０１０２】
各メディアオブジェクトテーブルは、メディアオブジェクトＩＤ、名称、ディスクリプション、ピクチャスタンプ、ＰＩＤ、フォーマット、スケジュールＩＤ、スクリプトＩＤ、マスタＩＤを含む多数のフィールドを有している。各メディアオブジェクトテーブルに関連して、ピープルテーブル、マスタテーブル、スケジュールテーブル、ストーリーボードテーブル、スクリプトテーブル、及び多数のショットテーブルがある。
【０１０３】
マスタテーブルは、メディアオブジェクトテーブルのマスタＩＤフィールドに関連付けられたマスタＩＤ、タイトル、ベーシックＵＭＩＤ、ＥＤＬ ＩＤ、テープＩＤ、継続時間、ピクチャスタンプを含む多数のフィールドを有している。
【０１０４】
スケジュールテーブルは、スケジュールＩＤ、スケジュール名、ドキュメントフォーマット、パス、作成日、原作者、開始日、終了日、バージョン、最終変更日、変更者、ノート、プログラムＩＤに関連付けられたＰＩＤを含む多数のフィールドを有している。
【０１０５】
契約テーブルは、プロダクションスタッフテーブルとキャストテーブルと団体テーブルの契約ＩＤに関連付けられた契約ＩＤ、開始日、レート、ジョブタイトル、満了日、詳細を含んでいる。
【０１０６】
ストーリーボードテーブルは、ショットテーブルのストーリーボードに関連付けられたストーリーボードＩＤ、ディスクリプション、著者、パス、メディアＩＤを含む多数のフィールドを有している。
【０１０７】
各ショットテーブルは、ショットＩＤ、ＰＩＤ、メディアＩＤ、タイトル、ロケーションＩＤ、ノート、ピクチャスタンプ、スクリプトＩＤ、スケジュールＩＤ、ディスクリプションを含む多数のフィールドを有している。各ショットテーブルに関連して、ピープルテーブル、スケジュールテーブル、スクリプトテーブル、ロケーションテーブル、及び、多数のテイクテーブルがある。
【０１０８】
ロケーションテーブルは、ショットテーブルのロケーションＩＤフィールドに関連付けられたロケーションＩＤ、ＧＰＳ、所在地、ディスクリプション、名称、１時間毎の費用、ディレクション、連絡先名、連絡先住所、連絡先電話番号、ピクチャスタンプを含む多数のフィールドを有している。
【０１０９】
各テイクテーブルは、ベーシックＵＭＩＤ、テイク番号、ショットＩＤ、メディアＩＤ、タイムコード・イン、タイムコード・アウト、サインメタデータ、テープＩＤ、カメラＩＤ、ヘッド時間、ビデオグラファ、インスタンプ、アウトスタンプ、レンズＩＤ、オートＩＤ、取り込みＩＤ、ノートを含む多数のフィールドを有している。各テイクテーブルに関連して、テープテーブル、タスクテーブル、カメラテーブル、レンズテーブル、取り込みテーブル、及び、多数のテイクアノテーションテーブルがある。
【０１１０】
取り込みテーブルは、テイクテーブルにおける取り込みＩＤに関連付けられた取り込みＩＤと、ディスクリプションとを含んでいる。
【０１１１】
テープテーブルは、テイクテーブルのテープＩＤフィールドに関連付けられたテープＩＤ、ＰＩＤ、フォーマット、最大継続時間、第１用途、最大削除、現削除、ＥＴＡ（予測到着時刻）、最終削除日、ピクチャスタンプを含む多数のフィールドを有している。
【０１１２】
タスクテーブルは、タスクＩＤ、ＰＩＤ、テイクテーブルのメディアＩＤ及びショットＩＤフィールドに関連付けられたメディアＩＤ及びショットＩＤ、タイトル、タスクノート、分配リスト、ＣＣリストを含む多数のフィールドを有している。タスクテーブルに関連して、プランドショットテーブルがある。
【０１１３】
プランドショットテーブルは、プランドショットＩＤ、タスクテーブルのＰＩＤ、メディアＩＤ、ショットＩＤにそれぞれ関連付けられたＰＩＤ、メディアＩＤ、ショットＩＤ、ディレクター、ショットタイトル、ロケーション、ノート、ディスクリプション、ビデオグラファ、期日、プログラムタイトル、メディアタイトル、アスペクト比、フォーマットを含む多数のフィールドを有している。
【０１１４】
カメラテーブルは、テイクテーブルのカメラＩＤフィールドに関連付けられたカメラＩＤ、製造者、モデル、フォーマット、シリアル番号、ヘッド時間、レンズＩＤ、ノート、連絡先名、連絡先住所、連絡先電話番号、ピクチャスタンプを含む多数のフィールドを有している。
【０１１５】
レンズテーブルは、テイクテーブルのレンズＩＤフィールドに関連付けられたレンズＩＤ、製造者、モデル、シリアル番号、連絡先名、連絡先住所、連絡先電話番号、ピクチャスタンプを含む多数のフィールドを有している。
【０１１６】
各テイクアノテーションテーブルは、テイクアノテーションＩＤ、ベーシックＵＭＩＤ、タイムコード、シャッタースピード、アイリス、ズーム、ガンマ、ショットマーカＩＤ、フィルタホイール、詳細及び利得を含む多数のフィールドを有している。各テイクアノテーションテーブルに関連して、ショットマーカテーブルがある。
【０１１７】
ショットマーカテーブルは、テイクアノテーションテーブルのショットマーカＩＤに関連付けられたショットマーカＩＤと、ディスクリプションとを含む多数のフィールドを有している。
【０１１８】
上述の説明では、ビデオマテリアルを例として本発明を説明したが、本発明は、ビデオマテリアル、オーディオマテリアル、データマテリアルのいずれか１以上に適用することができる。
【０１１９】
ＵＭＩＤは例えば３２〜６４ビットと長い。ウォーターマークは、データベースに記憶されたＵＭＩＤに関する別の短いアイデンティファイアを有してもよい。このように、ウォーターマークが基づくコードは間接的にマテリアルを識別する。
【０１２０】
本発明の実施例ではデジタル信号を使用しているので、プログラマブルデジタル信号プロセッサを用いて本発明を行うことができる。従って、本発明には、デジタル信号プロセッサで動作する場合に本発明を行うコンピュータプログラム製品と、このようなコンピュータプログラム製品を記憶する記憶媒体も含まれる。
【図面の簡単な説明】
【図１】 本発明にかかるシステムを示す概略ブロック図である。
【図２】 ウォーターマーキングシステムを示す概略ブロック図である。
【図３】 ウォーターマーキングシステムを示す概略ブロック図である。
【図４】 シグネチャ生成及び検出システムを示す概略ブロック図である。
【図５】 シグネチャ生成及び検出システムを示す概略ブロック図である。
【図６】 マテリアルアイデンティファイアのデータ構造を示す図である。
【図７】 マテリアルアイデンティファイアのデータ構造を示す図である。
【図８】 メタデータベースにおけるデータ構造例を示す図である。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to material identification. The present invention also relates to a method and apparatus for identifying such a material. Thus, the material is represented by an information signal, preferably a digital signal.
[0002]
The material may be one or more of video material, audio material, and data material. The video material will be described below for convenience of explanation. However, the present invention is not limited to video material. The present invention can also be applied to audio material. Further, the present invention can be applied to data material other than audio and video.
[0003]
[Prior art]
It is known to identify video material by adding “watermarks” to the video signal. A watermark is an encoded signal and is combined with a video signal. The watermark can be combined in such a way that the encoded signal is not visible or hardly visible in the displayed image. The encoded signal can be detected in the video signal. This is used, for example, to detect piracy. However, the watermark may be intentionally or accidentally damaged or removed by further processing of the material, for example by editing or adding special effects.
[0004]
DISCLOSURE OF THE INVENTION
The present invention provides a material identification method comprising the steps of inserting an identification code into a material as a watermark and extracting a signature from the material.
[0005]
The step of retrieving the signature may be before or after the step of inserting the watermark.
[0006]
In a preferred embodiment of the present invention, the signature is data taken from the data characteristics of the material or from the data characteristics of the material. The signature is extracted from the material. For example, the signature may be the position of the maximum and minimum values of the signal. Also, for example, the signature may be the coordinates of the brightest and darkest pixels of each frame in the video sequence. The signature may also be a histogram of measurable components of the material. For example, it may be a histogram of chrominance values of a video sequence. Signatures vary by material. The signature is retrieved from the material without the need to add data to the material.
[0007]
The present invention also provides an apparatus having a code generator for generating an identification code, a watermark generator for generating a watermark from the code and combining the watermark with a material, and a signature generator for extracting a signature from the material. To do.
[0008]
The watermark is vulnerable to damage and / or removal from the video material. If the watermark is damaged or removed, it becomes difficult or impossible to prove the origin of the material. The properties of the material that generates the signature can also be damaged or removed, but not the same process as the removal or damage of the watermark. Taking a signature that is a property of the material and having a watermark provides two independent means of proving its origin. Thus, if one side is damaged or removed, the other can be used to prove the origin.
[0009]
By providing a code as a watermark on the material, the material retains an ID in a series of productions and distributions after the code is provided. Further, a separate channel such as an identification code tape track is not required. Thereby, subsequent material processing becomes easy by associating and holding the identification code with the material.
[0010]
In the preferred method, this code identifies the material instead of or in addition to identifying ownership.
[0011]
The code is preferably used to associate the material with the metadata associated with the material.
[0012]
More preferably, the signature and identification code are stored in a database. Monitor the material to determine if a watermark is included. The monitored material is processed to extract the signature. If a watermark exists, the watermark and retrieved signature are compared with the watermark and signature in the database to determine the origin of the material.
[0013]
In one embodiment of the present invention, the material is recorded on an example of a recording medium, and the identification code is a code that identifies an example of the recording medium. The identification code may be a serial number, for example. The identification code may be used to refer to a UMID (unique material identifier).
[0014]
In other embodiments, the code is a UMID.
[0015]
Since the UMID and the code for identifying an example of the recording medium are shorter than the conventional watermark code, they can withstand the processing of the material and are difficult to remove.
[0016]
In one embodiment where the code is UMID, the instance number of the UMID is used to give each UMID and a unique watermark to each individual user or group of users for multiple copies of a material piece.
[0017]
BEST MODE FOR CARRYING OUT THE INVENTION
For better understanding of the present invention, embodiments of the present invention will be described below with reference to the accompanying drawings.
[0018]
Hereinafter, the present invention will be described using video materials. Note that the present invention is not limited to video material.
[0019]
Overview
Referring to FIG. 1, video source 880 provides video material. The source may be a camera, video tape recorder, video server, or other suitable source. A watermark generator 882 inserts or adds a watermark to the video material. Watermarking will be described later in the “Watermarking” section. In the preferred embodiment of the present invention, the watermark is generated from the UMID. The UMID is generated by a UMID generator 886. The UMID will be described later in the “UMID” section. Thus, in the preferred embodiment, a watermark based on UMID identifies the material and the material owner can also be identified. The UMID can identify a material with an accuracy of one frame. Thus, in the preferred embodiment, a watermark exists for each frame and uniquely identifies each frame. In another preferred embodiment, the instance number field of the UMID is used as a “fingerprint” that distinguishes various authorized users of the material.
[0020]
Video material from the source is also fed to a signature generator 884 that extracts the signature from the material. The signature will be described later in the “Signature” section.
[0021]
A database processor 894 is also provided. This is connected to the UMID generator 886 and the signature generator 884 via suitable communication links such as cables, RF links, computer network links, and the like. Database 894 stores the UMID and signature generated by generators 884 and 886. Database 894 associates the signature retrieved from the video material with the UMID attached to the same material.
[0022]
The material with the watermark is distributed or transmitted to the user via the distribution / transmission network 888. A monitoring station 890 monitors the network 888 to detect signatures and watermarks in the video material. For example, the material is broadcast. Station 890 includes a receiver and a video recorder that records the received video. The station also has the same signature generator 891 as the generator 884. The station also has a watermark detector 893. The received video is processed at the station and the signature is extracted from the material by exactly the same process as that performed by the generator 884. Also, the received video is processed and if there is a watermark, it is taken out and the watermark UMID is detected.
[0023]
The signature and UMID extracted from the received material are compared with the signature and UMID in the database 894 (892). The material distributed in this way can be confirmed and, for example, misuse can be detected. Further, since the UMID uniquely identifies the material, for example, royalty can be collected by the UMID.
[0024]
It is advantageous to have both a signature and a watermark, as watermarks can be removed or at least damaged in the distribution network, especially when material is processed by editing or applying special effects. There is also a possibility of intentionally removing the watermark. Similarly, video material may be modified by a process that reduces the likelihood of retrieving matching signatures from the original and processed material. By providing both, one can be used to confirm the origin of the material.
[0025]
Database 894 may include metadata associated with the material by UMID regarding the material. An example of metadata held by the database 894 will be described later in the “Metadata” section.
[0026]
Watermarking-Figures 2 and 3
The amount of information is increasing, and in particular, the amount of information on digitally recorded, stored, and distributed video is increasing. Such information is easy to duplicate, but is problematic because the copyright of the underlying work can be easily infringed by unauthorized copies. For this reason, the copyright holder must be able to show that there is no proper protection against the copy, or that the particular work sample was taken from the copyright holder's work and is an unauthorized copy that infringes the rights. For example, they don't like making copyrighted works available or distributed.
[0027]
One possible technique for obtaining a solution to this problem is digital watermarking. With digital watermarking, for example, a code including information for identifying an owner or a distributor and / or an authentication code can be embedded in a digital work. The digital watermark may be used with other deterrence means such as encryption.
[0028]
A digital watermark is referred to below as a watermark, but the watermark must be unique, for example, to clearly identify the owner or distributor and / or provide an authentication code. This is a technique often called fingerprinting. The watermark itself may be a digital work such as an image, audio, or video. The watermark can also include an indication that the work can be copied freely, cannot be copied at all, or can be copied a predetermined number of times.
[0029]
Preferably, the watermark should be such that it cannot be detected, changed or removed by unauthorized persons. Also, the watermark should not degrade the base work prominently. However, the watermark must be easily recognized by an authorized person so that the owner and / or distributor can be identified.
[0030]
The watermark must be able to be easily embedded in the underlying digital work. Preferably, the embedding technique can be easily performed during recording, thereby adding a watermark to the work at the source, minimizing the possibility of using a work without the watermark.
[0031]
The watermark can be placed, for example, in the header or label of the digital work, or it can be embedded in the data field of the digital work itself. Preferably, the watermark is played back many times in the work, more preferably present in each frame of the digital work. Alternatively, the watermark may be placed directly on the medium having the digital work.
[0032]
The watermark can be robust so that a person attempting an unauthorized copy cannot be removed or degraded. If the robust watermark is illegally removed, the data is seriously deteriorated and the data cannot be used. When data includes a lot of extra information such as a video, a robust watermark is likely to be attacked by, for example, a frame drop. Thus, a robust watermark should preferably be able to withstand such attacks, and may use error correction / recovery techniques that change from frame to frame and apply to data, for example.
[0033]
Also, the watermark may be fragile so as to be damaged if unauthorized copying is performed.
[0034]
However, the watermark should preferably be reversible and removable by the owner if necessary. The removal of the watermark is particularly effective in the post-production stage, for example, in order to reduce the cumulative effect of the watermark of the base work. In addition, when editing information from various sources, it is desirable to add another watermark to the edited product.
[0035]
The end user equipment can be configured to recognize the watermark so as not to allow copying of the protected work. In addition, the device can be configured to operate only on a work obtained from a specific owner and distributed via a specific distributor, or only a work having a specific authentication code.
[0036]
The watermark can be taken out by comparing the data with the watermark with the data without the watermark and confirming the validity.
[0037]
Two methods for embedding a watermark in the data field of a digital work are described in detail below. The first method embeds the watermark in the spatial domain, and the second method embeds the watermark in the frequency domain. Neither embedding process should cause significant degradation of the watermarked data.
[0038]
Spatial area watermark
Generally speaking, in this process, a predetermined data bit is changed by a watermark bit for creating watermarked data. The presence of the watermark can be determined by performing a cancellation operation on the data with the watermark.
[0039]
One approach is to embed the watermark by replacing the lower bits of the pseudo-randomly selected data with bits representing the watermark. However, these watermarks are easily destroyed by processing the least significant bit of data. Another approach is to insert a geometric pattern representing the watermark into the data. However, these watermarks are easily destroyed by data geometric processing. As yet another approach, the watermark is embedded in a manner similar to quantization noise. This will be described later with reference to FIG. 5, and the document “Embedding Secret Information into a Dithered Multi-Level Image” K Tanaka et al., IEEE Military Communications Conference, 1990, pp.216-220 and the document “Video Steganography”. K Mitsui, IMA Intellectual Property Proceedings, volume 1, 1994, pp.187-296 has a detailed explanation. However, these watermarks are easily destroyed by signal processing, particularly data requantization.
[0040]
Referring now to FIG. 2, source 650 generates a digital data signal 652 such as digital video. The watermark inserter 700 is connected to the source 650 and receives the digital data signal 652. The watermark inserter 700 adds a watermark to the digital data signal 652 by a method similar to re-quantization noise, thereby adding a watermark 663 to generate data 705 with the watermark. The storage device 670 is connected to the watermark inserter 700 and stores data 705 with a watermark.
[0041]
As yet another approach, n pairs of image points (ai, bi) are randomly selected to increase the brightness of ai by 1 and decrease the brightness of bi by 1. If certain statistical properties of the image are met, the sum of the differences between the n pairs of points is 2n.
[0042]
Alternatively, when the data signal is composed of at least two components (for example, [Y, U, V] according to MPEG, PAL, and NTC), watermarks can be embedded by assigning values that do not normally occur in combination to these components. For example, when embedding a watermark in video data having two image fields, a positive watermark is arranged in the first field and a negative watermark is arranged in the second field. When the image field is reproduced, there is a masking effect due to the interlace of the field, and the visibility of the watermark is greatly reduced.
[0043]
Frequency domain watermark
Generally speaking, this process obtains a frequency spectrum image of the data to be watermarked. The watermark is embedded in a predetermined component of the frequency spectrum image. After that, the frequency spectrum image with the watermark is subjected to inverse transformation to generate data with the watermark. The watermark can be extracted by performing a cancellation operation on the data with the watermark.
[0044]
One approach is to divide the data into blocks and compute the discrete cosine transform (DCT) for each of these blocks. Thereafter, the frequency coefficient of each block is adjusted. A pseudo-random subset of blocks is selected, and in each such block, the frequency coefficients of the predetermined subset are adjusted such that the relative value encodes the data bits. The variation of the relative value and the selection of a predetermined subset of frequencies must be made insensitive to the watermark. This watermark is susceptible to damage from noise and further processing.
[0045]
Also, as described in detail in European Patent Application 0766468 by NEC, watermarks can be encoded by fine-tuning each frequency coefficient. This has the advantage of making the watermark less susceptible to damage, but increases the overall noise level.
[0046]
Referring now to FIG. 3, source 650 generates a digital data signal 652 such as digital video. The frequency converter 655 is connected to the source 650 and receives the digital data signal 652. The frequency converter 655 converts the digital data signal 652 into the frequency spectrum data 657 using, for example, discrete cosine transform or fast Fourier transform. The watermark inserter 660 is connected to the frequency converter and receives the frequency spectrum data 657. The watermark inserter adds the watermark 663 by adjusting each coefficient of the frequency spectrum data 657, and generates the frequency spectrum data 663 with the watermark. The inverse frequency converter 665 is connected to the watermark inserter 660 and receives the frequency spectrum data 663 with the watermark. The inverse frequency converter 665 converts the frequency spectrum data 663 with a watermark into data 667 with a watermark. The storage device 670 is connected to the inverse frequency converter 665 and stores data 667 with a watermark.
[0047]
Yet another approach is to use the presence of the so-called masking phenomenon in the human visual and auditory system to increase the change for a particular frequency coefficient. Masking occurs when an area of data is disturbed by a perceptually more prominent area of the data. However, these areas must be identified before the insertion of the watermark, which increases the complexity of embedding.
[0048]
As yet another approach, digital data is compressed and watermarks are embedded in the x and y coordinates of the motion vector of the compressed data. This is more robust to processing because it has the advantage that the watermark is embedded after compression.
[0049]
Signature-FIGS. 4 and 5
The amount of information is increasing, and in particular, the amount of information on digitally recorded, stored, and distributed video is increasing. Such information is easy to duplicate, but is problematic because the copyright of the underlying work can be easily infringed by unauthorized copies. For this reason, the copyright holder must be able to show that there is no proper protection against the copy, or that the particular work sample was taken from the copyright holder's work and is an unauthorized copy that infringes the rights. For example, they don't like making copyrighted works available or distributed.
[0050]
One technique for solving this problem is to use a so-called digital signature. A digital signature can be generated for any piece of digital information, such as an image. Digital signatures of works that are suspected of infringement are also generated using the same technique. The correlation between the two signatures shows that the suspicious work is a copy of the original work.
[0051]
For digital work such as audio or video, for example, a sequence of digital signatures can be generated from individual frames of data, for example. A sequence of digital signatures for a work that is suspected of being infringed can also be generated using the same technique. The correlation between two consecutive signature sequences shows that the suspicious work is a copy of the original work.
[0052]
The digital signature may be used with other deterrent means such as watermarking or encryption.
[0053]
The digital signature is hereinafter referred to as a signature, but is preferably easily generated from the work. Preferably, the signature is generated by applying an algorithm in the work.
[0054]
Preferably, the algorithm generates a signature related to the empirical nature of the work. Preferably, the algorithm should be robust so that a high correlation between the signature generated from the original work and the signature generated from the copy work can be maintained. Preferably, the algorithm should be robust to easily implemented processing techniques and noise introduction.
[0055]
The signature of the original work can be generated and stored for later reference. Preferably, an automated system can be used to monitor a data stream containing suspicious work. The automation system generates suspicious work signatures, compares them to the stored original work signatures, and alerts the original work owner if there is a correlation.
[0056]
The end-user equipment may be configured to recognize the signature of the original work so as not to allow copying of the original work. Further, the device may be configured to play back only works having a predetermined signature.
[0057]
Hereinafter, the signature generation process will be described in detail with reference to FIG. Source 850 generates a digital data signal 852, such as an original digital video work. Signature generator 855 is connected to source 850 and receives digital data signal 852. Signature generator 855 generates signature 857 by applying algorithm 856 to digital data signal 852. Storage device 860 is connected to signature generator 855 and stores signature 857 associated with digital data signal 852.
[0058]
Next, the process of determining whether or not the suspicious work is a copy of the original work will be described in detail with reference to FIG. Source 865 generates a digital data signal 867, such as an original digital video work. Signature generator 855 is connected to source 865 and receives digital data signal 867. Signature generator 855 generates signature 858 by applying algorithm 856 to digital data signal 867. The signature comparator 870 is connected to the signature generator 855 and the storage device 860. The storage device 860 has a signature of the original digital data work. Signature comparator 870 receives signature 858 and compares the signature in storage 860 with signature 858 and generates an alarm signal 873 if there is a correlation. Alarm 875 is connected to signature comparator 870 and generates an alarm upon receipt of alarm signal 873.
[0059]
Next, two possible algorithms for signature generation are described in detail.
[0060]
The first algorithm generates a signature by determining the position of the feature in the work. The digital data signal is analyzed to determine, for example, the position of the maximum and minimum values of the signal. The generated signature has the position of the maximum and minimum values of the signal, either by the spatial coordinate system or the bit number of the data stream. Preferably, in a digital video signal, these maximum and minimum values correspond to the highest and lowest chrominance values of a frame of pixels, and the generated signature has the coordinates of these pixels.
[0061]
The second algorithm generates a signature by determining a histogram of the work. The digital data signal is analyzed to determine the measurable component of the signal. The generated signature has histogram coefficients representing these measurable components. Preferably, in the digital video signal, the measurable component is a chrominance value or luminance value of one frame, and the generated signature has a histogram coefficient representing the chrominance value or luminance value of the frame.
[0062]
UMID—FIGS. 6 and 7
UMID is described in reference [2]. FIG. 6 shows the extended UMID. It has a first set of 32 bytes consisting of basic UMID and a second set of 32 bytes consisting of signature metadata.
[0063]
The first set of 32 bytes is a basic UMID. The following are the components.
[0064]
A 12-byte universal label for identifying this first set as SMPTE UMID. This defines the type of material that the UMID identifies, as well as how to create a globally unique material number and a locally unique instance number.
A 1-byte length value that defines the remaining length of the UMID.
A 3-byte instance number used to distinguish between different material instances with the same material number.
A 16-byte material number used to identify each clip. Each material number is the same for related instances of the same material.
[0065]
A second set of 32 bytes of signature metadata as a set of packed metadata items is used to create an extended UMID. In the extended UMID, signature metadata comes immediately after the basic UMID, and the signature metadata includes the following components.
[0066]
An 8-byte date / time code that identifies the date / time when the content unit was created
A 12-byte value that determines the spatial coordinates when creating the content unit.
• Three 4-byte codes for registering country, group, and user codes.
[0067]
The basic and extended UMID components will be described below.
[0068]
"12-byte universal label"
The first 12 bytes of the UMID identify the UMID with a registered string value defined in Table 1.
[0069]
[Table 1]

[0070]
Hexadecimal values in Table 1 may be changed. These values are examples. The first to twelfth bytes may be specified other than those shown in the table. In the example shown in Table 1, the fourth byte indicates that the fifth to twelfth bytes relate to a data format that matches SMPTE. The fifth byte indicates that the sixth to tenth bytes relate to “dictionary” data. The sixth byte indicates that such data is “metadata” defined by the seventh to tenth bytes. The seventh byte indicates a part of the dictionary including the metadata defined by the ninth and tenth bytes. The eighth byte indicates the dictionary version. The ninth byte indicates the class of data, and the tenth byte indicates a specific item in the class.
[0071]
In this embodiment, the first to tenth bytes have a fixed value assigned in advance. The eleventh byte is a variable. Referring to FIG. 7 and Table 1 above, it can be seen that the first to tenth bytes of the UMID label are fixed. Thus, these bytes can be replaced by a one-byte “type” code T representing the first through tenth bytes. A length code L follows the type code T. Following this, there are 2 bytes, one of which is the 11th byte of Table 1 and the other is the 12th byte of Table 1, an instance number (3 bytes), and a material number (16 bytes). In addition, optionally, the signature metadata of extended UMID and / or other metadata may be arranged after the material number.
[0072]
The UMID type (11th byte) has four values for identifying four different data types as follows.
[0073]
“01h” = UMID of picture material
“02h” = UMID of audio material
“03h” = UMID of data material
“04h” = UMID of group material (for example, combination of related contents)
The last (12th) byte of the 12-byte label identifies how to create the material and instance number. This byte is divided into a top nibble and a bottom nibble, where the top nibble defines the material number creation method and the bottom nibble defines the instance number creation method.
[0074]
"length"
The length is a 1-byte number having a value “13h” for the basic UMID and a value “33h” for the extended UMID.
[0075]
"Instance number"
The instance number is a unique 3-byte number created by any of several means defined in the standard. The instance number provides a link between a particular “instance” of a clip and external related metadata. Without this instance number, all materials are tied to any instance of the material and its associated metadata.
[0076]
Creating a new clip requires creating a new material number with an instance number of zero. Thus, a non-zero instance number indicates that the associated clip is not the source material. Instance numbers are primarily used to identify relevant metadata about a particular instance of a clip.
[0077]
"Material number"
The 16-byte material number is a non-zero number that can be created by any of several means defined in the standard. This number varies depending on the 6-byte registered port ID number, time, and random number generator.
[0078]
"Signature Metadata"
If any component from the signature metadata cannot contain a meaningful value, a null fill can be performed. Any null-filled component is completely null-filled to clearly indicate to the subsequent decoder that the component is not valid.
[0079]
"Date and time format"
The date and time format is 8 bytes, and the first 4 bytes are UTC (Universal Time Code) based on the time component. The time is defined by either the AES3 32-bit audio sample clock or SMPTE12M, depending on the type of content.
[0080]
The latter 4 bytes define data based on modified Julian data (MJD) as defined in SMPTE 309M. This counts 999,999 days from midnight November 17, 1858, allowing dates up to 4597 AD.
[0081]
"Spatial coordinate format"
A spatial coordinate value consists of three components defined as follows.
-Altitude: 8 decimal numbers specifying up to 99,999,999 meters.
Longitude: Eight decimal numbers that specify 180.000 degrees east / west longitude (valid to the fifth decimal place).
Latitude: Eight decimal numbers specifying 90.000 degrees north latitude / south latitude (valid to the fifth decimal place).
[0082]
Since the altitude value is expressed in meters from the center of the earth, altitudes below the sea level are also possible.
[0083]
Note that the spatial coordinates are static for most clips, but not in all cases. In the case of materials captured from a moving source such as a camera attached to a vehicle, the spatial coordinate values may change.
[0084]
"Country code"
The country code is an abbreviated 4-byte alphanumeric string according to the set defined by ISO 3166. Unregistered countries can obtain a registered alphanumeric string from the SMPTE Registration Authority.
[0085]
“Group Code”
The organization code is an abbreviated 4-byte alphanumeric string registered in SMPTE. Since organization codes are meaningful only with respect to their registered country codes, organization codes can have the same value in different countries.
[0086]
"user code"
The user code is a 4-byte alphanumeric string assigned locally by each organization but not globally registered. Since user codes are defined only for their registered organization codes and country codes, user codes can have the same value in different organizations and countries.
[0087]
"Freelance operator"
The freelance operator uses his country of residence as the country code, for example a group code and user code combined with an 8-byte code that can be registered with SMPTE. These freelance codes have a 7 digit alphanumeric string that begins with the "~" symbol (ISO 8859 character number 7Eh) and is subsequently registered.
[0088]
From the above description, it can be seen that UMID can be used to identify not only video material, but also audio material, data material, and material group.
[0089]
Metadata-Figure 8
Hereinafter, the types of metadata that can be generated during the creation of a program and the configuration approach that can be used to configure the metadata will be described using examples. The UMID embedded in the video as a watermark serves as an identifier that links the video to the associated metadata in the database.
[0090]
FIG. 8 shows an example of a structure for configuring metadata. A number of tables comprising a number of fields each having metadata are provided. These tables can be related to each other by a common field in each table to obtain a relational structure. This structure may also have multiple instances of the same table to represent multiple instances of the object that the table represents. The field is formatted in a predetermined way. The field size can also be predetermined. Examples of the size include “Int” representing 2 bytes, “LongInt” representing 4 bytes, and “Double” representing 8 bytes. Also, the size of the field can be determined with reference to the number of characters held in the field, such as 8, 10, 16, 32, 128, 255 characters.
[0091]
A detailed description of this structure is a program table. The program table has a number of fields including a program ID (PID), a title, a working title, a genre ID, a list, an aspect ratio, a director ID, and a picture stamp. In relation to the program table, there are a genre table, a keyword table, a script table, a people table, a schedule table, and a plurality of media object tables.
[0092]
The genre table has a number of fields including a genre ID associated with the genre ID field of the program table and a genre description.
[0093]
The keyword table has a number of fields including a program ID associated with the program ID field of the program table, a keyword ID, and a keyword.
[0094]
The script table has a number of fields including script ID, script name, script type, document format, path, creation date, original author, version, last modified date, modified person, PID associated with the program ID, and notes. ing. The people table has a number of fields containing images.
[0095]
The people table is associated with a number of individual tables and a number of group tables. Each individual table has a number of fields containing images. Each group table has a number of fields containing images. Each individual table is associated with either a production staff table or a cast table.
[0096]
Production Staff Table includes Production Staff ID, Last Name, First Name, Contract ID, Agent, Agency ID, Email, Address, Phone Number, Role ID, Notes, Allergy, Birthday, National Insurance Number, Bank ID, Picture Stamp It has a number of fields to include.
[0097]
The cast table has a number of fields including cast ID, last name, first name, title, contract ID, agent, agency ID, equity number, email, address, phone number, birthday, bank ID, picture stamp. Yes. Related to production staff tables and cast tables are bank detail tables and agency tables.
[0098]
The bank detail table has a number of fields including a bank ID, a sort code, an account number, and an account name associated with the bank ID field of the production staff table and the bank ID field of the cast table.
[0099]
The agency table has a number of fields including the agency ID field associated with the agency ID field in the production staff table and the agency ID field in the cast table, name, location, phone number, website and email, and picture stamp. . In addition, there is a role table related to the production staff table.
[0100]
The role table has a number of fields including a role ID, job, note, and picture stamp associated with the role ID field of the production staff table. Each group table is associated with a group table.
[0101]
The organization table has a number of fields including organization ID, name, type, location, contract ID, contact name, contact phone number, website, and picture stamp.
[0102]
Each media object table has a number of fields including media object ID, name, description, picture stamp, PID, format, schedule ID, script ID, and master ID. Associated with each media object table is a people table, a master table, a schedule table, a storyboard table, a script table, and a number of shot tables.
[0103]
The master table has a number of fields including a master ID, title, basic UMID, EDL ID, tape ID, duration, and picture stamp associated with the master ID field of the media object table.
[0104]
The schedule table has a number of fields including schedule ID, schedule name, document format, path, creation date, original author, start date, end date, version, last modified date, modifier, note, PID associated with the program ID have.
[0105]
The contract table includes a contract ID, a start date, a rate, a job title, an expiration date, and details associated with the contract IDs of the production staff table, cast table, and organization table.
[0106]
The storyboard table has a number of fields including the storyboard ID, description, author, path, and media ID associated with the shot table storyboard.
[0107]
Each shot table has a number of fields including a shot ID, a PID, a media ID, a title, a location ID, a note, a picture stamp, a script ID, a schedule ID, and a description. Associated with each shot table is a people table, a schedule table, a script table, a location table, and a number of take tables.
[0108]
The location table includes the location ID, GPS, location, description, name, hourly cost, direction, contact name, contact address, contact phone number, picture stamp associated with the location ID field of the shot table. It has a number of fields to include.
[0109]
Each take table consists of basic UMID, take number, shot ID, media ID, time code in, time code out, signature metadata, tape ID, camera ID, head time, videographer, instamp, outstamp, lens It has a number of fields including ID, auto ID, capture ID and notes. Associated with each take table is a tape table, task table, camera table, lens table, capture table, and a number of take annotation tables.
[0110]
The capture table includes a capture ID associated with the capture ID in the take table and a description.
[0111]
The tape table includes a tape ID, PID, format, maximum duration, first usage, maximum deletion, current deletion, ETA (predicted arrival time), last deletion date, and picture stamp associated with the tape ID field of the take table. It has many fields.
[0112]
The task table has a number of fields including task ID, PID, media ID and shot ID associated with the take table media ID and shot ID field, title, task note, distribution list, and CC list. In connection with the task table, there is a planned shot table.
[0113]
The plan shot table is a plan shot ID, a task table PID, a media ID, a PID associated with each shot ID, a media ID, a shot ID, a director, a shot title, a location, a note, a description, a videographer, a date, a program It has a number of fields including title, media title, aspect ratio, format.
[0114]
The camera table is the camera ID, manufacturer, model, format, serial number, head time, lens ID, notes, contact name, contact address, contact phone number, picture stamp associated with the camera ID field of the take table Has a number of fields.
[0115]
The lens table has a number of fields including lens ID, manufacturer, model, serial number, contact name, contact address, contact phone number, picture stamp associated with the lens ID field of the take table. .
[0116]
Each take annotation table has a number of fields including take annotation ID, basic UMID, time code, shutter speed, iris, zoom, gamma, shot marker ID, filter wheel, details and gain. Associated with each take annotation table is a shot marker table.
[0117]
The shot marker table has a number of fields including a shot marker ID associated with the shot marker ID of the take annotation table and a description.
[0118]
In the above description, the present invention has been described using video material as an example. However, the present invention can be applied to any one or more of video material, audio material, and data material.
[0119]
The UMID is as long as 32 to 64 bits, for example. The watermark may have another short identifier for the UMID stored in the database. Thus, the code on which the watermark is based indirectly identifies the material.
[0120]
Since embodiments of the present invention use digital signals, the present invention can be implemented using a programmable digital signal processor. Accordingly, the present invention includes a computer program product that implements the present invention when operating on a digital signal processor and a storage medium that stores such a computer program product.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram showing a system according to the present invention.
FIG. 2 is a schematic block diagram showing a watermarking system.
FIG. 3 is a schematic block diagram showing a watermarking system.
FIG. 4 is a schematic block diagram illustrating a signature generation and detection system.
FIG. 5 is a schematic block diagram illustrating a signature generation and detection system.
FIG. 6 is a diagram illustrating a data structure of a material identifier.
FIG. 7 is a diagram illustrating a data structure of a material identifier.
FIG. 8 is a diagram illustrating an example of a data structure in a meta database.

Claims (35)

Inserting the identification code into the material as a watermark;
Extracting the signature from the material by analyzing the position of at least one of the maximum and minimum values of the signal characterizing the material or analyzing the histogram of the chrominance or luminance values of the material;
Storing the identification code and signature in a database;
Examining the material for the presence of the watermark, and if so, comparing the material's watermark to the watermark in the database;
Extract more signatures from the material by analyzing the position of at least one of the maximum and minimum values of the signal that characterizes the material, or analyzing the histogram of the chrominance or luminance values of the material, and the extracted signatures in the database A material identification method comprising: comparing with a signature of. The method of claim 1, wherein the identification code identifies the material directly or indirectly. 3. A method according to claim 1 or 2 , wherein metadata about the material is stored in a database and the identification code associates the material with the metadata in the database. The material has a frame,
Identification code any one method of claims 1-3, characterized in that present in each frame of the material. Any one method of claims 1-4, characterized in that it comprises an identification code a time code. Identification code any one method of claims 1-5, characterized in that to identify the owner of the material. Identification code indirectly method of any one of claims 1-6, characterized in that to identify the material. 8. The method of claim 7 , wherein the material is recorded on a recording medium and the identification code identifies an example of the medium on which the material is recorded. Any one method of claims 1-6 identification code, characterized in that identifying the direct material. The method of claim 9 , wherein the identification code uniquely identifies the material. The method of claim 10 , wherein the identification code is a UMID. The method of claim 11 , wherein an instance number field of the UMID is used to include the fingerprinting code. Identification code, any one method according to claim 1 to 12, characterized in that it comprises data identifying the authorized user of the material. The method of claim 13 , wherein if the identification code is a UMID, data identifying an authorized user is included in the instance number field of the UMID. Material is video material, audio material, any one method according to claim 1-14, characterized in that at any one or more of the data material. Watermarks method towards any one of claims 1 to 15, characterized in that one of a frequency domain watermark and a time domain watermark. Method towards any one of claims 1 to 16, wherein the watermark is robust and / or revocable. A code generator for generating an identification code;
A watermark generator that generates a watermark from the identification code and combines the watermark with the material;
A signature generator that extracts the signature from the material by analyzing the position of at least one of the maximum and minimum values of the signal that characterizes the material, or analyzing the histogram of the chrominance or luminance values of the material;
A database processor for storing the identification code and the signature retrieved by the signature generator ;
Inspect the material for the presence of the watermark , and characterize the material from the material by analyzing the location of at least one of the signal's maximum and minimum values, or analyzing the histogram of the material's chrominance or luminance values Take out the monitor,
And a processor for comparing the watermark of the material with the watermark in the database and comparing the signature retrieved by the monitor with the signature in the database. The apparatus of claim 18, wherein the identification code directly or indirectly identifies the material. The database processor stores metadata about materials ,
Apparatus have a communication link for connecting to the database processor,
20. An apparatus according to claim 18 or 19 , wherein the identification code associates material with metadata in the database. The material has a frame,
21. The apparatus of any one of claims 18-20 , wherein the generator is configured to insert an identification code into each frame of material. The apparatus according to any one of claims 18 to 21 , wherein the identification code includes a time code. 23. A device according to any one of claims 18 to 22 , wherein the identification code identifies the owner of the material. 24. An apparatus according to any one of claims 18 to 23 , wherein the identification code indirectly identifies the material. The apparatus has a recorder for recording material on a recording medium,
The apparatus of claim 24 , wherein the identification code identifies an example of a medium on which material is recorded. 24. A device according to any one of claims 18 to 23 , wherein the identification code directly identifies the material. 27. The apparatus of claim 26, wherein the identification code uniquely identifies the material. 28. The apparatus of claim 27 , wherein the identification code is UMID. 29. The apparatus of claim 28, wherein a UMID instance number field is used to contain the fingerprinting code. 30. A device according to any one of claims 18 to 29 , wherein the identification code comprises data identifying an authorized user of the material. 31. The apparatus of claim 30 , wherein if the identification code is a UMID, data identifying an authorized user is included in the instance number field of the UMID. 32. The apparatus according to any one of claims 18 to 31 , wherein the material is one or more of a video material, an audio material, and a data material. Watermark, equipment of any one of claims 18 to 32, characterized in that one of a frequency domain watermark and a time domain watermark. Equipment of any one of claims 18-33, wherein the watermark is robust and / or revocable. When operating in a digital signal processor, the computer program for performing the method of any one of claims 1 to 17.

Images