JP3851689B2 - Image reconstruction processing device - Google Patents

Description

【００７８】
次に、ステップ５４( ＳＴ５４ )の処理として、この重み付き補間されたデータに逆投影の重みＷｂを乗算する。
【００７９】
ここで、ステップ５５( ＳＴ５５ )の処理として、この逆投影の重みを乗算したデータをメモリ２６の該当ピクセルに対応する位置( アドレス )に加算する。
次に、ステップ５６( ＳＴ５６ )の処理として、ピクセル数カウンタのカウント値が０であるか否かを判断する。ここで、ピクセル数カウンタのカウント値が０ではないと判断すると、ステップ５７( ＳＴ５７ )の処理として、ピクセル数カウンタのカウント値に対して−１の減算処理を行い、逆投影処理を次のピクセルへ移行させ、再び前述のステップ５３の処理へ戻るようになっている。
【００８０】
また、ピクセル数カウンタのカウント値が０であると判断すると、１ビューにおける傾斜断層面画像の１列( Ｚ座標が同じ )への逆投影処理が終了したことになるので、この各逆投影部の逆投影処理を終了するようになっている。
この各逆投影部の逆投影処理を終了すると、再び前述の傾斜画像再構成処理に戻るようになっている。
【００８１】
このようにこの第３の実施の形態によれば、前述の第２の実施の形態と同様な効果を得ることができる。
なお、上述したステップ５３の処理及びステップ５４の処理において、図１５に示すように、同一回転における位相の異なる２つのデータに対するチャンネル方向のＷｈの重み付き補間Ｃ１，Ｃ２を並列処理により行い、これらの並列処理Ｃ１，Ｃ２により得られたデータに対してＺ軸方向の( ヘリカルスキャンに相当する )Ｗｒの重み付き補間Ｃ12をパイプライン処理で接続すれば、処理の高速化をさらに向上させることができる。
なお、この第３の実施の形態においてもマルチスライスＣＴ装置においても適用できるものである。
【００８２】
この発明の第４の実施の形態を図１６乃至図２３を参照して説明する。なお、この第４の実施の形態も、前述の各実施の形態と同様に、前述の第１の実施の形態における傾斜画像再構成処理の内容が異なるだけで、基本的なハードウエアの構成は同一であるので、同一部材には同一符号を付してその説明は省略する。 この第４の実施の形態においては、図１６に示す傾斜画像再構成処理の流れに基づいて、再構成処理制御部２１が傾斜断層面画像を直接再構成する。
【００８３】
まず、ステップ６１( ＳＴ６１ )の処理として、予備的な撮影として行われたスキャノグラム撮影により得られた画像から設定された傾斜断層面画像をＺ座標の単位で分解する。
次に、ステップ６２( ＳＴ６２ )の処理として、データ収集部１１で収集され、各種補正が行われた検出データに対して、コンボリューション処理を行う。
【００８４】
次に、ステップ６３( ＳＴ６３ )の処理として、後述するヘリカル補間用の重みＷｒによる逆投影処理を行い、このＷｒによる逆投影処理を終了すると、ステップ６４( ＳＴ６４ )の処理として、( １−Ｗｒ )による逆投影処理を行い、この( １−Ｗｒ )による逆投影処理を終了すると、この傾斜画像再構成処理を終了するようになっている。
【００８５】
図１７は、前述したステップ６２の処理のＷｒによる逆投影処理及び前述したステップ６３の処理の( １−Ｗｒ )による逆投影処理の流れを示す図である。
( １−Ｗｒ )による逆投影処理は、Ｗｒによる逆投影処理中のＷｒが( １−Ｗｒ )に置き換わるだけである。
まず、ステップ６５( ＳＴ６５ )の処理として、１次元Ｘ線検出器５のビュー数( 例えば１０００ )をデータメモリ２２に形成されたビュー数カウンタにセットし、ステップ６６( ＳＴ６６ )の処理として、前述の傾斜画像再構成処理のステップ６１の処理で分解された傾斜断層面画像のＺ座標範囲から導かれる逆投影する列数をデータメモリ２２に形成された列数カウンタにセットする。
【００８６】
これらの各カウンタへのセットが終了すると、ステップ６７( ＳＴ６７ )の処理として、次の４列のピクセルデータに対応する、２回転分のコンボリューションデータと共に各４列のピクセルデータに逆投影するためのデータ( 逆投影開始データ位置、データ回転位置、逆投影ピッチ( チャンネルピッチ )、データ回転ピッチ( Ｚ座標方向の１回転の移動量 )、逆投影開始点、計算回数( ピッチ数等 ) )を各逆投影部２３-1〜２３-4のそれぞれのメモリ２６へ転送する。
【００８７】
次に、ステップ６８( ＳＴ６８ )の処理として、後述する各逆投影部のＷｒ( 又は１−Ｗｒ )による逆投影処理を行い、このＷｒ( 又は１−Ｗｒ )による逆投影処理を終了すると、ステップ６９( ＳＴ６９ )の処理として、各逆投影部２３-1〜２３-4のそれぞれのメモリ２６に記憶された１列の画像データを画像加算部２５を介して画像メモリの該当する列位置に加算する。
【００８８】
次に、ステップ７０( ＳＴ７０ )の処理として、列数カウンタのカウント値が０か否かを判断する。ここで列数カウンタのカウント値が０ではないと判断すると、ステップ７１( ＳＴ７１ )の処理として、列数カウンタのカウント値に対して−４の減算処理を行い、逆投影処理を次の４列へ移行させ、再び前述のステップ６７の処理へ戻るようになっている。
【００８９】
また、列数カウンタのカウント値が０であると判断すると、１ビューにおける傾斜断層面画像( ピクセル )への逆投影処理が終了したことになるので、ステップ７２( ＳＴ７２ )の処理として、ビュー数カウンタのカウント値が０であるか否かを判断する。ここで、ビュー数カウンタのカウント値が０ではないと判断すると、ステップ７３( ＳＴ７３ )の処理として、ビュー数カウンタのカウント値に対して−１の減算処理を行い、逆投影処理を次のビューへ移行させ、再び前述のステップ６６の処理へ戻るようになっている。
【００９０】
また、ビュー数カウンタのカウント値が０であると判断すると、全ビューにおける傾斜断層面画像( ピクセル )への逆投影処理が終了したことになるので、このＷｒ( 又は１−Ｗｒ )による逆投影処理処理を終了するようになっている。このＷｒ( 又は１−Ｗｒ )による逆投影処理処理を終了すると、再び前述の傾斜画像再構成処理へ戻るようになっている。
【００９１】
図１８は、前述したステップ６８の処理において各逆投影部２３-1〜２３-4で行われる各投影部のＷｒ( 又は１−Ｗｒ )による逆投影処理の流れを示す図である。なお、ここでは第１の逆投影部２３-1で行われる逆投影処理について説明し、他の各逆投影部２３-2〜２３-4で行われる逆投影処理も同様であるので、その説明は省略する。
まず、ステップ７４( ＳＴ７４ )の処理として、列数カウンタのカウント値及びメモリ２６に記憶されたデータに基づいて、メモリ２６に形成されたピクセル数カウンタに該当する列のピクセル数( 計算回数 )をセットする。
【００９２】
次に、ステップ７５( ＳＴ７５ )の処理として、メモリ２６に記憶された逆投影開始データチャンネル位置Ｄｈｊ、逆投影ピッチΔ、データ回転ピッチに基づいて、逆投影するピクセルに対して２個のコンボリューションデータ( 所定の回転数目の該当するビューのｋチャンネル目及びｋ＋１チャンネル目のデータＤｋ、Ｄ( ｋ＋１ )を選択し、この２個のデータに対してデータアドレス重み発生器２８により発生させたチャンネル方向の補間用の重みＷｈで重み付き補間を行う。 ステップ７６( ＳＴ７６ )の処理として、この補間したデータに対して、所定の逆投影重みＷｂとヘリカル補間の重みＷｒ( 又は１−Ｗｒ )との積を乗算する。
なお以下に示す、式( ６ )はチャンネル方向の補間用の重みＷｈの算出式、式( ７ )は式( ６ )におけるＮｈｋの算出式、式( ８ )は逆投影データＤｐの算出式の一例である。
【００９３】
【数２】

【００９４】
次に、ステップ７７( ＳＴ７７ )の処理として、この逆投影重みとヘリカル補間の重みを乗算したデータをメモリ２６の該当するピクセルに対応する位置( アドレス )に加算する。
【００９５】
次に、ステップ７８( ＳＴ７８ )の処理として、ピクセル数カウンタのカウント値が０であるか否かを判断する。ここで、ピクセル数カウンタのカウント値が０ではないと判断すると、ステップ７９( ＳＴ７９ )の処理として、ピクセル数カウンタのカウント値に対して−１の減算処理を行い、逆投影処理を次のピクセルへ移行させ、再び前述のステップ７５の処理へ戻るようになっている。
【００９６】
また、ピクセル数カウンタのカウント値が０であると判断すると、１ビューにおける傾斜断層面画像の１列( Ｚ座標が同じ )への逆投影処理が終了したことになるので、この各逆投影部のＷｒ( 又は１−Ｗｒ )による逆投影処理を終了するようになっている。この各逆投影部のＷｒ( 又は１−Ｗｒ )による逆投影処理を終了すると、再び前述のＷｒ( 又は１−Ｗｒ )による逆投影処理に戻るようになっている。
【００９７】
従って、図１６に示す傾斜画像再構成処理の流れのように、Ｗｒによる逆投影処理の次に( １−Ｗｒ )による逆投影処理を行うようになっており、これによりヘリカル補間が達成される。
このようにこの第４の実施の形態によれば、前述の第２の実施の形態と同様な効果を得ることができる。
【００９８】
なお、この第４の実施の形態では、逆投影処理を２回行って傾斜断層面画像の再構成を行うようになっていたが、この発明はこれに限定されるものではなく、図１９に示すように、ヘリカル補間の重みＷｒｉ( ｉ＝１，…，ｎ、Ｗｒ１＋Ｗｒ２＋…＋Ｗｒｎ＝１ )による逆投影処理をｎ回行って傾斜断層面画像を再構成するものでも良いものである。なお、図１９に示す傾斜画像再構成処理は、図１６に示す傾斜画像再構成処理における内容( 図１７及び図１８参照 )と基本的には同一であり、ヘリカル補間の重みＷｒの取扱いのみが異なるので、ここでは、その詳細な説明は省略する。
【００９９】
なお、ヘリカル補間の重みＷｒ( Ｗｒｉ )により２回( ３回以上 )逆投影処理を行うようになっているが、これらのヘリカル補間の重みＷｒによる各逆投影処理を並列処理するようにして、その並列処理の後で画像加算すれば計算時間の短縮が可能となる。
【０１００】
ところで、ヘリカル補間の重みＷｒを、図２０に示すａのように傾斜断層面画像をＺ座標にしたがって複数に分割した各領域に対して該当する領域のみ１となり、その他の領域で０となるように重み付けすれば、傾斜断層面画像の再構成で各領域が別々に再構成されるので、前述の第１の実施の形態を実現することができる。
【０１０１】
しかし、これを図２０に示すｂのように傾斜断層面画像をＺ座標にしたがって複数個に分割した各領域に対して、該当する領域では１、その両隣りの領域では１次関数的に減衰するように重み付けすれば、第１の実施の形態において、１枚の傾斜断層面画像における各領域間の段差( ぎざぎざ )を改善して滑らかな１枚の傾斜断層面画像が得られる。
【０１０２】
また、図２１に示すように、所定のビュー( ｎ )におけるヘリカル補間の重みａを、１回転( １回転はＮビュー )前の同位相のビュー( ｎ−Ｎ )においてヘリカル補間の重みｂの範囲とオーバーラップしない部分では１とし、オーバーラップする部分では１から０に１次関数的に減衰するように重み付けし、図２２に示すように、所定のビュー( ｎ )の重みａと隣接するビュー( ｎ＋１ )における重みｂは、傾斜断層面画像においては単なる平行移動となるので、隣接するビュー( ｎ＋１ )における画像再構成では、所定のビュー( ｎ )の再構成結果を流用することができる。なお、所定のビュー( ｎ )と、１回転前の同位相のビュー( ｎ−Ｎ )と、隣接するビュー( ｎ＋１ )のヘリカルスキャンにおける位置関係は、図２３に示すようになっている。
【０１０３】
また、ヘリカル補間の重みＷｒは、単純な形式の数値であるため、テーブル等に記憶しておき、計算時に呼出すようにしても良い。さらに、Ｗｒは、再構成するピクセル単位で変化しても良いし、またピクセル列単位で変化しても良いものである。
さらに、この第４の実施の形態では、シングルスライスＣＴ装置のヘリカルスキャンについて説明したが、この発明はこれに限定されるものではなく、マルチスライスＣＴ装置のスタティックスキャンのものに適用することができる。
【０１０４】
この発明の第５の実施の形態を図２４乃至図２６を参照して説明する。この第５の実施の形態は、図２４に示すように、隣接する傾斜断層面画像を複数枚再構成することを特徴とするものである。隣接する傾斜断層面画像どうしでは、重複して使用できるデータが存在するので、効率的に画像再構成を図ることが可能となる。なお、この第５の実施の形態も、前述の各実施の形態と同様に、前述の第１の実施の形態における傾斜画像再構成処理の内容が異なるだけで、基本的なハードウエアの構成は同一であるので、同一部材には同一符号を付してその説明は省略する。
【０１０５】
まず、隣接する傾斜断層面画像のピッチを適切に調節すると、図２５( ａ )〜( ｄ )に示すように、隣接する各傾斜断層面画像をＺ軸に垂直に分割した時にそれぞれ傾斜断層面画像がどれも同じように分割されるようにすることができる。
隣接する傾斜断層面画像を複数枚再構成する第１の方法は、この図２５( ａ )〜( ｄ )に示すように分割して、前述の第１の実施の形態のように、それぞれの傾斜断層面画像を各分割領域毎に制限して逆投影処理を行うものである。この時、各分割領域について逆投影するとき使用する逆投影開始データ位置、逆投影ピッチ、逆投影開始点( 加算開始点 )、計算回数等の情報が共通となり、これらの情報を取扱う計算処理を短縮することができる。
【０１０６】
また第２の方法は、第１の逆投影部２３-1〜第４の逆投影部２３-4において、それぞれが担当する傾斜断層面画像１枚ずつ受持って逆投影処理を行うのではなく、図２６に示すように、各逆投影部２３-1〜２３-4が複数枚の傾斜断層面画像のＺ座標に垂直に分割された同じスライス位置の部分を受持つようにする。
そうすることにより、各逆投影部２３-1〜２３-4が補間データを重複して使用することなく、１つの逆投影部が１スライス位置の補間データを使用して、複数枚の傾斜断層面画像にわたって存在するその補間データが逆投影される部分は全部一貫して逆投影処理することになる。
【０１０７】
各逆投影部２３-1〜２３-4がそれぞれ専用の補間データを使用して逆投影した各分割領域のデータは、画像メモリ２４上で加算合成され、それぞれ１枚の傾斜断層面画像となる。図２６中の傾斜断層面画像の１１、２１、３１、…、５３、６２、７１は、例えば１１は、データ( 補間データ )１を逆投影して得た分割領域１のデータであることを示している。
【０１０８】
従って、この第２の方法によれば、傾斜断層面画像毎に分割領域の個数分の補間データを作成する必要がなく、しかも、データメモリ２２から各補間データを該当する逆投影部２３-1〜２３-4のメモリの内の１つのメモリへ転送するだけで良いので、効率的に画像再構成を図ることが可能となる。
このようにこの第５の実施の形態によれば、隣接する傾斜断層面画像を複数枚再構成する場合に処理時間を短縮することができる効率的な画像再構成を図ることが可能となる。
【０１０９】
この発明の第６の実施の形態を図２７乃至及び図３１を参照して説明する。この第６の実施の形態では、マルチスライスＣＴにおいてスタティックスキャンを行う場合について説明する。なお、この第６の実施の形態も、前述の各実施の形態と同様に、前述の第１の実施の形態における傾斜画像再構成処理の内容が異なるだけで、基本的なハードウエアの構成は同一であるので、同一部材には同一符号を付してその説明は省略する。なお、この第６の実施の形態では、４列マルチスライスＣＴを使用した例で説明する。
【０１１０】
この第６の実施の形態においては、図２７に示す傾斜画像再構成処理の流れに基づいて、再構成処理制御部２１が傾斜断層面画像を直接再構成する。
まず、ステップ８１( ＳＴ８１ )の処理として、予備的な撮影として行われたスキャノグラム撮影により得られた画像を表示装置１４に表示して、設定操作入力に基づいて傾斜断層面画像を設定し、この設定した傾斜断層面画像を例えば図２８乃至図３０に示すように、Ｚ座標に垂直に４列のＸ線ビームのビーム厚にしたがって分割する。
【０１１１】
すなわち、図２８に示すように、ＦＯＶ( 撮影領域 )に対して４列のＸ線ビームが放射され、このＸ線ビームのＦＯＶの中心軸に占める範囲( 各厚さｔ )が、図２９( ａ )に示すように、分割するＺ軸( ＦＯＶの中心軸と一致する )に垂直な各スライスの位置( 各厚さｔ )となる。
図２９( ａ )に示すように、４枚のスライスにわたって傾斜断層面画像が設定された時には、図２９( ｂ )に示すように、４個の領域に分割される。また、図３０( ａ )に示すように、３枚のスライスにわたって傾斜断層面画像が設定された時には、図３０( ｂ )に示すように、３個の領域に分割される。
【０１１２】
このようにして傾斜断層面画像の分割が終了すると、ステップ８２( ＳＴ８２ )の処理として、架台を回転しＸ線を照射して、１回転分のデータ収集を行う。この収集されたデータに対してはＸ線強度補正及び検出器感度補正等の各種補正を行う。
このデータ収集を終了すると、ステップ８３( ＳＴ８３ )の処理として、ビュー数カウンタにカウント値１を設定して、ステップ８４( ＳＴ８４ )の処理として、１回目の再構成か否かを判断する。
【０１１３】
ここで、１回目の再構成と判断すると、ステップ８５( ＳＴ８５ )の処理として、ビュー数カウンタに設定されている数値に該当するビューの４列のデータ( ステップ８２の処理で得られたデータ )をそれぞれ再構成関数とコンボリューションして、４列のコンボリューションデータを算出する。
次に、ステップ８６( ＳＴ８６ )の処理として、各コンボリューションデータを各逆投影部２３-1〜２３-4に割当てて転送し、各逆投影部２３-1〜２３-4によりそれぞれ分割領域へ限定した逆投影処理を行い、ステップ８７( ＳＴ８７ )の処理として、各逆投影部２３-1〜２３-4での逆投影処理で得られた分割領域の画像を画像メモリ２４上で加算して１枚の傾斜断層面画像を作成し、表示装置１４に表示する。
【０１１４】
次に、ステップ８８( ＳＴ８８ )の処理として、ビュー数カウンタのカウント値が１回転分のビュー数Ｎに等しいか否かを判断する。
ここで、ビュー数カウンタのカウント値が１回転分のビュー数Ｎに等しくない、すなわちＮ未満と判断すると、ステップ８９( ＳＴ８９ )の処理として、ビュー数カウンタのカウント値に対して＋１の加算更新処理を行い、この＋１の加算更新処理を終了すると、再び前述のステップ８６の処理へ戻るようになっている。
【０１１５】
また、ビュー数カウンタのカウント値が１回転分のビュー数Ｎに等しいと判断すると、表示装置１４には必要なビュー数( 例えば１８０°分でも良い )で再構成された正確な傾斜断層面画像が表示されている。ここで、ステップ９０( ＳＴ９０ )の処理として、傾斜断層面画像の更新を行うか否かを判断する。
ここで、傾斜断層面画像の更新を行うと判断すると、再び前述のステップ８２の処理へ戻るようになっている。このことにより、前述のステップ８４の処理において、１回目の再構成ではなく、２回目以上の再構成と判断するようになる。
【０１１６】
前述のステップ８４の処理で、１回目の再構成ではなく、２回目以上の再構成と判断すると、ステップ９１( ＳＴ９１ )の処理として、同位相( 同一ビュー )・同検出器列における前回のデータ収集で得られたデータ( 補正されたデータ )と今回のデータ収集で得られたデータ( 補正されたデータ )との差を算出して差分データを得る。
【０１１７】
次に、ステップ９２( ＳＴ９２ )の処理として、得られた差分データをそれぞれ再構成関数とコンボリューションして差分のコンボリューションデータを得る。この差分コンボリューションデータを得ると、前述のステップ８６の処理へ移行して、各分割領域毎に差分コンボリューションデータが逆投影され、画像メモリ２４で加算され、差分コンボリューションデータによる前回の傾斜断層面画像の更新( 画像の変化部分だけを変更する )が行われる。
【０１１８】
前述のステップ９０の処理で、傾斜断層面画像の更新は行わないと判断すると、ステップ９３( ＳＴ９３ )の処理として、傾斜断層面画像の変更を行うか否かを判断する。
ここで、傾斜断層面画像の変更を行うと判断すると、再び前述のステップ８１の処理へ戻るようになっている。
また、傾斜断層面画像の変更を行なわないと判断すると、この傾斜画像再構成処理を終了するようになっている。
【０１１９】
このようにこの第６の実施の形態によれば、マルチスライスＣＴのスタティックスキャンにおいて、前述した第１の実施の形態と同様な効果を得ることができる。
なお、この第６の実施の形態において、前述した第５の実施の形態の技術を使用して隣接する傾斜断層面画像を再構成しても良いものであり、また前述した第４の実施の形態等の技術を使用して傾斜断層面画像を再構成しても良いものである。
【０１２０】
また、実効スライス厚の厚い画像を得るためには、例えば図３１に示すように、１０列の２次元Ｘ線検出器を使用した場合、各検出器列のデータを予め加算( 束ね処理 )しておけば良い。
すなわち、第１の検出器列に対応するスライス位置においては、第１の検出器列の各セグメントのデータを２倍にしたものにそれぞれ第２の検出器列の同一位置のセグメントのデータを加算したものを使用し、第２の検出器列に対応するスライス位置においては、第１の検出器列、第２の検出器列、第３の検出器列の各セグメントのデータをそれぞれ加算したものを使用し、第３の検出器列に対応するスライス位置においては、第２の検出器列、第３の検出器列、第４の検出器列の各セグメントのデータをそれぞれ加算したものを使用し、…、第１０の検出器列に対応するスライス位置においては、第９の検出器列の各セグメントのデータに第１０の検出器列の同一位置のセグメントのデータを２倍にしたものを加算したものを使用すれば良い。
なお、Ｘ線コーンビームの外側でデータのない領域については、隣接するデータを逆投影して近似するか、又は全く逆投影せずに黒色で表示しても良い。
【０１２１】
この発明の第７の実施の形態を図３２、図３３、図４３及び図４４を参照して説明する。
この第７の実施の形態は、円錐状に広がったＸ線( コーンビーム )に対向してＸ線検出素子を平面的( ２次元的 )・マトリックス状に配列した２次元Ｘ線検出器を前述の実施の形態の１次元Ｘ線検出器５の代わりに配置したコーンビームＣＴにおける任意の傾斜断層面画像を再構成するものである。
【０１２２】
コーンビームＣＴにおける２次元画像及び３次元画像の再構成法については、特願平８−１０１５号に詳細に記載されている。この特願平８−１０１５号には、図４３に示すように、Ｘ線源２０１と２次元Ｘ線検出器２０２との間の撮影領域ＦＯＶにおいてボクセル２０３を設定し、図４４に示すように、このボクセル２０３のボクセル列( 例えば２０３-1 )に平行なセンタリング列が配列されたセンタリング面２０４を設定し、２次元Ｘ線検出器２０２から得られたデータを補間処理等を行った後、一度センタリング面２０４に逆投影し、このセンタリング面２０４からボクセル２０３へ逆投影して２次元画像又は３次元画像を再構成するもので、このような再構成法により、再構成処理時間の大幅な短縮を実現することができる。
【０１２３】
また、この特願平８−１０１５号のセンタリング面に一度逆投影した後に各ボクセル列に逆投影する投影法を使用して再構成することにより、任意の傾斜断層面画像をＸ線データから直接再構成することは可能である。
すなわち、センタリング面からボクセルへの逆投影において、各ボクセル列において傾斜断層面画像に該当するボクセルだけに制限すれば、上述した第２の実施の形態と同様に、直接再構成することができる。また、ボクセルを制限したときにセンタリング面も制限することができ、２次元Ｘ線検出器からセンタリング面への逆投影においても、該当するセンタリング面に制限すれば、逆投影処理及びその結果として再構成処理にかかる時間をより短縮することができる。
【０１２４】
また、コーンビーム照射のヘリカルスキャンにおいては、図３２に示すように傾斜断層面画像として２焦点のビームから得られたデータを逆投影して再構成されるものでは、図３３に示す傾斜画像再構成処理の流れのように、傾斜断層面画像を各ビームに対応した領域に分割し、その分割領域毎に領域制限した２回の逆投影処理により再構成すれば良い。
【０１２５】
すなわち、まず、ステップ１０１( ＳＴ１０１ )の処理として、位相( ビュー )を格納する位相エリアθに０°を設定し、ステップ１０２( ＳＴ１０２ )の処理として、傾斜断層面画像を位相エリアθにより設定された位相における各焦点位置に対応してＺ軸に垂直に分割する。
【０１２６】
次に、ステップ１０３( ＳＴ１０３ )の処理として、回転数カウンタＮに初期値をセットし、ステップ１０４( ＳＴ１０４ )の処理として、回転数カウンタＮにセットされている数値に対応するＮ回転目の分割領域に位相エリアθにセットされている位相に対応する位相θで収集されたデータにより領域限定のコーンビーム逆投影を行い、表示装置１４に逆投影により再構成された画像を表示する。この領域限定のコーンビーム逆投影を終了すると、ステップ１０５( ＳＴ１０５ )の処理として、領域限定のコーンビーム逆投影が位相θについて全ての焦点で終了したか否かを判断する。すなわち、位相θについて傾斜断層面画像全体へのコーンビーム逆投影が終了したか否かを判断することになる。
【０１２７】
ここで、領域限定のコーンビーム逆投影が位相θについて全ての焦点で終了していないと判断すると、ステップ１０６( ＳＴ１０６ )の処理として、回転数カウンタＮのカウント値に対して＋１の加算更新処理を行い、再び前述のステップ１０４の処理へ戻るようになっている。
【０１２８】
また、領域限定のコーンビーム逆投影が位相θについて全ての焦点で終了したと判断すると、ステップ１０７( ＳＴ１０７ )の処理として、必要な位相( ３６０°分又は１８０°分 )のコーンビーム逆投影が全て終了したか否かを判断する。
ここで、必要な位相のコーンビーム逆投影がまだ終了していないと判断すると、ステップ１０８( ＳＴ１０８ )の処理として、位相エリアθの設定値に対して＋Δθの加算更新処理を行い、再び前述のステップ１０２の処理へ戻るようになっている。
また、必要な位相のコーンビーム逆投影が全て終了したと判断すると、この傾斜画像再構成処理を終了するようになっている。
【０１２９】
このようにこの第７の実施の形態によれば、コーンビームＣＴにおいても傾斜断層面画像をボクセルデータを作成した後で再構成するのではなく、ボクセルデータを介さずに２次元Ｘ線検出器から収集されたデータから直接再構成することができる。
また、３６０°補間法によりヘリカル補間を行って１回転分( ３６０°分 )のデータを作成するものについて説明したが、この発明はこれに限定されるものではなく、例えば対向ビーム補間法、対向ビームを使用したフィルタ補間法、隣接補間法、隣接補間法等の各種の補間法を使用しても良いものである。
【０１３０】
さらに、１回転分( ３６０°＋α )のデータによる再構成法を使用したが、この発明はこれに限定されるものではなく、半回転( ハーフ、１８０°＋α )のデータによる再構成法を使用しても良いものである。
また、Ｘ線検出器についても、均等８列検出器による４列サンプリングでも良いし、直線マルチスライス検出器あるいは平面検出器でも良く、その他の形状のＸ線検出器によるデータ収集でも良い。
なお、この発明は上述した実施の形態に限定されるものではなく、この発明の要旨を逸脱しない範囲で変形することができる。
【０１３１】
【発明の効果】
以上詳述したようにこの発明によれば、医師が所望とする傾斜断層面画像を複数の領域に分割して再構成処理を行うので、傾斜断層面画像の再構成処理に係る時間を短くすることができる。
【図面の簡単な説明】
【図１】この発明の第１の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の概略の構成を示すブロック図。
【図２】同実施の形態のＸ線ＣＴ装置のジオメトリを示す図。
【図３】同実施の形態のＸ線ＣＴ装置の再構成処理部の構成を示すブロック図。
【図４】同実施の形態のＸ線ＣＴ装置の再構成処理部の逆投影部の構成を示すブロック図。
【図５】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図６】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における傾斜断層面画像の分割を説明する図。
【図７】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における各分割領域の画像データの画像メモリ上での加算を説明する図。
【図８】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における逆投影処理の流れを示す図。
【図９】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における逆投影開始データ位置、逆投影開始点及び逆投影ピッチを説明する図。
【図１０】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における必要とする逆投影領域を示す図。
【図１１】この発明の第２の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図１２】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における各逆投影部の逆投影処理の流れを示す図。
【図１３】この発明の第３の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図１４】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における各逆投影部の逆投影処理の流れを示す図。
【図１５】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における重み付き補間の並列処理及びパイプライン処理を説明するための模式図。
【図１６】この発明の第４の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図１７】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理におけるＷｒ( 又は１−Ｗｒ )による逆投影処理の流れを示す図。
【図１８】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理におけるＷｒ( 又は１−Ｗｒ )による逆投影処理における各逆投影部のＷｒ( 又は１−Ｗｒ )による逆投影処理の流れを示す図。
【図１９】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の他の一例の流れを示す図。
【図２０】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における第１の実施の形態を実施するための重み付けの例を示す図。
【図２１】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における滑らかな傾斜断層面画像を再構成するための重み付けを示す図。
【図２２】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理における図２１の重み付けの隣接するビューでの流用を説明するための図。
【図２３】同実施の形態のＸ線ＣＴ装置の所定のビューと１回転前の同位相のビューと隣接するビューのヘリカルスキャンにおける位置関係を示す図。
【図２４】この発明の第５の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理において再構成する複数枚のそれぞれ隣接する傾斜断層面画像を示す図。
【図２５】同実施の形態のＸ線ＣＴ装置の傾斜画像再構成処理において再構成する複数枚のそれぞれ隣接する傾斜断層面画像の分割状態を示す図。
【図２６】同実施の形態のＸ線ＣＴ装置の傾斜画像再構成処理において複数枚のそれぞれ隣接する傾斜断層面画像の再構成における各逆投影部の受持ちを示す図。
【図２７】この発明の第６の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図２８】同実施の形態のＸ線ＣＴ装置におけるコーンビームの４列のビームを示す図。
【図２９】同実施の形態のＸ線ＣＴ装置におけるコーンビームの４列のビームと傾斜断層面画像及び傾斜断層面画像の分割の第１の例を示す図。
【図３０】同実施の形態のＸ線ＣＴ装置におけるコーンビームの４列のビームと傾斜断層面画像及び傾斜断層面画像の分割の第２の例を示す図。
【図３１】同実施の形態のＸ線ＣＴ装置における１０列の２次元Ｘ線検出器を使用した場合の束ね処理について説明する図。
【図３２】この発明の第７の実施の形態の画像再構成方法を使用するＸ線ＣＴ装置のコーンビームの２つの焦点に対する傾斜断層面画像の分割を示す図。
【図３３】同実施の形態のＸ線ＣＴ装置の再構成処理部で行われる傾斜画像再構成処理の流れを示す図。
【図３４】Ｘ線ＣＴ装置の１次元Ｘ線検出器及び２次元Ｘ線検出器を示す図。
【図３５】Ｘ線ＣＴ装置の２次元Ｘ線検出器の列方向及びチャンネル方向を示す図。
【図３６】Ｘ線ＣＴ装置のスタティクスキャン方式及びヘリカルスキャン方式を説明するための図。
【図３７】Ｘ線ＣＴ装置のヘリカルスキャンにおけるビューの回転位相及びＺ軸位置を示す図。
【図３８】Ｘ線ＣＴ装置のヘリカルスキャンにおけるスライスプロファイルを示す図。
【図３９】従来のボクセルデータからＭＰＲ処理により切り出す断面画像を示す図。
【図４０】従来の傾斜断層面画像のピクセルとボクセルデータとの関係を示す図。
【図４１】従来の傾斜断層面画像のボクセルデータの置換による画像のギザギザ状態を説明するための図。
【図４２】腰椎のスキャノグラムにおける傾斜断層面画像の位置の指定状態を示す図。
【図４３】特願平８−１０１５号に記載されているコーンビームに対するボクセルを説明するための図。
【図４４】特願平８−１０１５号に記載されているボクセル列及びセンタリング面を示す図。
【符号の説明】
３…Ｘ線源、
５…Ｘ線検出器、
１１…データ収集部、
１２…再構成処理部、
２１…再構成処理制御部、
２２…データメモリ、
２３…逆投影ブロック、
２３-1〜２３-4…第１〜第４の逆投影部、
２４…画像メモリ、
２５…画像加算部、
２６…メモリ、
２７…逆投影演算部、
２８…データアドレス重み発生部。[0001]
BACKGROUND OF THE INVENTION
The present invention irradiates an object with X-rays emitted from an X-ray source, detects X-rays transmitted through the object with an X-ray detector, and based on detection data obtained from the X-ray detector. An image reconstruction processing device for reconstructing a tomographic image of the object In place Related.
[0002]
[Prior art]
A single slice CT (computed tomography) apparatus, as shown in FIG. 34 (a), irradiates a subject with an X-ray beam in a fan shape from an X-ray source 101, is disposed opposite to the X-ray source 101, and includes an X-ray detection element. X-rays that have passed through the subject are detected by a one-dimensional X-ray detector 102 configured by arranging N channels (N) in a fan shape, for example, 1000 channels in a row.
Then, the X-ray source 101 and the one-dimensional X-ray detector 102 maintain a positional relationship facing each other with the subject interposed therebetween, and rotate around the subject while performing K view, for example, 1000 views (1000 times), The data (X-ray projection image) is reconstructed based on the X-ray data obtained from the one-dimensional X-ray detector 102.
[0003]
As shown in FIG. 34 (b) or FIG. 35, the multi-slice CT apparatus irradiates a subject with an X-ray beam in a conical shape from an X-ray source 103, and a one-dimensional X-ray detector 102 for a single slice CT apparatus is provided. A two-dimensional X-ray detector 104, 105 having X-ray detector elements arranged on the inner surface of the cylinder so that M rows are stacked in a direction perpendicular to the slice plane (Z-axis direction) (N channel × M segment). The transmitted X-ray is detected. Note that the two-dimensional X-ray detector 104 shown in FIG. 34B is an example of four segments.
Similar to the single-slice CT apparatus, the X-ray source 103 and the two-dimensional X-ray detector 104 rotate the periphery of the subject while rotating the surroundings of the subject while maintaining the opposed positional relationship across the subject. Data is collected, and an image (X-ray projection image) is reconstructed based on the X-ray data obtained from the two-dimensional X-ray detector 104.
[0004]
As shown in FIG. 36A, in the single slice CT apparatus and the multi-slice CT apparatus, the X-ray source 101 (103) and the one-dimensional (two-dimensional) X-ray detector 102 (104) are the same around the object. In the static scan (single scan or conventional scan) method that circulates the circular orbit of FIG. 36 and the single slice CT apparatus as shown in FIG. 36 (b), the X-ray source 101 and the one-dimensional X-ray detector 102 are The periphery continuously spirals, that is, the X-ray source 101 and the one-dimensional X-ray detector 102 orbit around the same circular orbit around the subject and place the subject in synchronization with the rotation. A helical scan method in which the top plate (bed) moves along the body axis (axis perpendicular to the slice plane) is performed.
Examples of the fan beam reconstruction method and the cone beam reconstruction method are described in the prior art and embodiments of Japanese Patent Application Nos. 8-1015, 8-1213, and 8-10218. Yes.
[0005]
In the helical scan method, when a tomographic image of a predetermined slice plane is reconstructed, the detection data corresponding to the predetermined slice plane position is only one view as shown in FIG. In FIG. 37, the vertical axis represents the rotational phase, the horizontal axis represents the Z-axis (body axis) position, and the rotational phase 0 ° and the rotational phase 360 ° are the same phase. The trajectory between the X-ray source 101 and the one-dimensional X-ray detector 102 is shown as a straight line, but the detection data of each view exists as a plurality of points (number of views) located at a constant interval on this straight line.
[0006]
Therefore, for example, at the slice position A on the Z-axis, the detection data of the jth view corresponds (exactly, if the slice position is arbitrarily selected, the detection data of one view may not correspond).
[0007]
In general, in order to reconstruct one tomographic plane image, 360 ° (or 180 °) view detection data is required.
Accordingly, the detection data of a predetermined tomographic plane image is obtained by interpolation with the detection data of two views having the same rotation phase across the tomographic plane image (slice plane) to be reconstructed, and the detection data for 360 ° is obtained.
[0008]
For example, the detection data C of the slice position A of the rotation phase B ° is obtained by using the detection data B1 of the view of the rotation phase B ° in the previous rotation (first rotation) and the view of the view of the rotation phase B ° in the rear rotation (second rotation). It calculates | requires from detection data B2.
That is, from the ratio of the distance L1 between the Z-axis position of the detection data B1 and the slice position A and the distance L2 between the Z-axis position of the detection data B2 and the slice position A
C = (L1 · B2 + L2 · B1) / (L1 + L2)
Thus, detection data of the rotational phase B ° at the slice position A is obtained.
In addition to the 360 ° interpolation method described above, detection data at a predetermined slice position can be obtained by various interpolation methods such as a counter beam interpolation method.
[0009]
By the way, in the X-ray CT apparatus, the response (slice profile) of the X-ray detection in the slice direction (body axis direction, Z axis direction) is not accurately rectangular but is obtained in a broken shape such as a single peak shape. . For example, when a thin plate-shaped phantom with a thickness of 0.2 mm or 0.1 mm is used, the detection data detected due to blurring of X-rays radiated from the X-ray source is substantially trapezoidal, and is helical with an actual helical pitch t. A slice profile obtained by interpolating detection data obtained by scanning has a deformed shape like a single peak as shown in FIG.
In this slice profile, with respect to the maximum value I0 of the detected signal intensity, the effective slice thickness D, which is the width in the signal intensity of (1/2) · I0, is an important factor that affects the image quality.
[0010]
The effective slice thickness of the axial image (thin plate-shaped phantom image) obtained by interpolating the detection data obtained by the helical scan method is the two beams having the thickness t (the first rotation beam and the second rotation beam). When the distance (helical pitch) of the beam) is t, the measured value is about 1.4 t.
As shown in FIG. 39, from voxel data obtained by stacking image data (normal slice plane images perpendicular to the Z-axis) reconstructed at intervals of slice thickness t, cross-sectional images inclined with respect to those image data (inclination) Tomographic image) MPR processing for cutting out E is known.
[0011]
At this time, for example, as shown in the divided diagrams of FIGS. 40 (a) and 40 (b), in order to obtain one pixel of the inclined tomographic plane image, eight adjacent voxels having this pixel as the approximate center. There is a way to interpolate from the data. In the case of this interpolation method, since the effective slice thickness of the voxel data is about 1.4 t, the thickness of the inclined tomographic plane image becomes 2 t or more by interpolation.
[0012]
As a method without interpolation for obtaining an inclined tomographic plane image, as shown in FIG. 41, a method of simply replacing the voxel data closest to each pixel of the inclined tomographic plane image as it is. In this method, as shown in FIG. 41, the inclined tomographic plane image becomes a stepped jagged image.
Actually, for example, when observing an abnormality of the lumbar vertebra, for example, as described above, a method of obtaining a cross-sectional image perpendicular to the lumbar vertebra by MPR processing, and a scanogram (as shown in FIG. Scan the fluoroscopic image without the helical scan), specify the position of the tilted tomographic plane image perpendicular to the target lumbar vertebra from the fluoroscopic image from the side of the body, and then based on the specified tilted tomographic plane image position There is a method in which the tilt angle (mounting frame tilt angle) is appropriately controlled at the time of scanning, and the data of the tilted tomographic plane image having the necessary tilt at the target slice position is collected.
[0013]
In order to treat the affected area while observing the cross-sectional image, a technique called CT fluoroscopy is known in which scanning is continued while the gantry is continuously rotated, and the cross-sectional image (slice image perpendicular to the body axis) is reconstructed in near real time. Yes. This technique of CT fluoroscopy is used, for example, in the case of performing biopsy reliably by inserting a needle while observing a cross-sectional image.
[0014]
By the way, when the needle is inserted in the lung field, the ribs get in the way, and if the cross-sectional image is tilted with respect to the plane perpendicular to the body axis by avoiding the ribs, it is best to insert the needle at an angle. Since a cross-sectional image (slice image) reconstructed by CT fluoroscopy is observed, a needle is inserted along a slice plane perpendicular to the body axis. Since this is based on slice images, it is a safe and secure procedure but not the best procedure.
[0015]
[Problems to be solved by the invention]
Scanograms are used to specify the target tilted tomographic plane image position, and based on this specified tilted tomographic plane image position, the tilt angle is controlled appropriately and sequentially during actual scanning to collect tilted tomographic plane image data. However, there is a problem that the scan time is long and the burden on the subject and the doctor is large.
Further, the method of obtaining an inclined tomographic plane image by MPR processing has a problem that the effective slice thickness is increased and the obtained inclined tomographic plane image becomes a blurred image. In the method of obtaining an inclined tomographic plane image by approximating with neighboring voxel data, there is a problem that a jagged image is obtained. Furthermore, both the method of obtaining an inclined tomographic plane image by MPR processing and the method of obtaining an inclined tomographic plane image by approximating with neighboring voxel data can be obtained after creating the voxel data once, so that it exceeds the time for creating the voxel data. There was a problem that it took time.
[0016]
In addition, a desired oblique tomographic plane image cannot be obtained by CT fluoroscopy, and although a doctor can perform a safe and reliable treatment, there is a problem that the best treatment cannot be performed.
[0017]
Accordingly, an object of the present invention is to provide an image reconstruction method in an image reconstruction processing apparatus capable of obtaining an inclined tomographic plane image desired by a doctor in a short time with a thin effective slice thickness.
[0018]
[Means for Solving the Problems]
The present invention provides a rotation center axis An X-ray source that generates a fan-shaped X-ray fan beam or a conical X-ray cone beam and an X-ray detector composed of one or more rows of X-ray detection elements are continuously rotated, and the subject is Helical scan data is collected by moving the table to be placed along the rotation center axis. In the image reconstruction processing device, an inclined tomographic plane inclined with respect to the rotation center axis A plurality of regions along the rotation center axis Means to divide into Means for generating projection data for one rotation for each region from the helical scan data by helical interpolation, and a plurality of regions related to the plurality of regions based on the projection data for one rotation. Means for reconstructing an image by convolution processing and back projection processing; The plurality image From The image corresponding to the inclined fault plane Generate Means.
[0026]
DETAILED DESCRIPTION OF THE INVENTION
A first embodiment of the present invention will be described with reference to FIGS.
FIG. 1 is a block diagram showing a schematic configuration of an X-ray CT apparatus using an image reconstruction method to which the present invention is applied.
A gantry 1 as a projection data measuring system includes a X-ray source 3 that generates an X-ray bundle of fan-shaped fan beams and a one-dimensional array type one-dimensional X-ray array in which a plurality of X-ray detection elements are arranged in a row. The line detector 5 is accommodated. The X-ray source 3 and the one-dimensional X-ray detector 5 are mounted on the rotating ring 2 in a state of facing each other with a subject placed on a slide top plate of a bed 6.
The one-dimensional X-ray detector 5 is composed of a plurality of (1000 channels) detector elements arranged in a one-dimensional array (see FIG. 1A) and mounted on the rotating ring 2. Is done. Here, one detection element is defined to correspond to one channel.
[0027]
X-rays from the X-ray source 3 are exposed to the subject via the X-ray filter 4. X-rays that have passed through the subject are detected as electrical signals by the one-dimensional X-ray detector 5.
The X-ray controller 8 supplies a trigger signal to the high voltage generator 7. The high voltage generator 7 applies a high voltage to the X-ray source 3 at the timing of receiving the trigger signal. As a result, X-rays are emitted from the X-ray source 3.
[0028]
The gantry bed controller 9 controls the rotation of the rotating ring 2 of the gantry 1 and the slide of the slide top plate of the bed 6 in synchronization. A system controller 10 serving as a control center of the entire system is configured so that the X-ray controller 3 performs so-called continuous rotation (for example, helical scan) in which the X-ray source 3 moves along a spiral trajectory as viewed from the subject. The gantry bed controller 9 is controlled.
Specifically, the rotating ring 2 continuously rotates at a constant angular velocity, the slide top plate of the bed 6 moves at a constant speed, and continuously or intermittently from the X-ray source 3 at every constant angle. The line is exposed.
[0029]
The output signal from the one-dimensional X-ray detector 5 is amplified by the data acquisition unit 11 for each channel and converted into a digital signal. Projection data output from the data collection unit 11 is taken into the reconstruction processing unit 12.
The reconstruction processing unit 12 obtains back projection data reflecting the X-ray absorption rate for each voxel based on the projection data.
[0030]
In a continuous rotation type X-ray CT apparatus using a fan beam, the effective field of view (FOV, imaging region) has a cylindrical shape centered on the rotation center axis of continuous rotation. Voxels (three-dimensionally arranged pixels) are defined (see FIG. 39), and the back projection data of each voxel is obtained from the projection data from the one-dimensional X-ray detector 5. The three-dimensional image data or tomographic image data created based on the backprojection data is sent to the display device 14 and visually displayed as a three-dimensional image or tomographic image.
[0031]
As shown in FIG. 2, the geometry of this X-ray CT apparatus is
Number of detector rows M = 1 segment,
Number of channels N = 1000 channels,
Slice thickness tmm
Focus-rotation center distance FCD (Focus-center-Distance) = 600mm,
Focus-detector distance FDD (Focus-Detector-Distance) = 1200 mm,
Effective field diameter FOV (Field of View) = 500mm,
Effective viewing angle (fan angle) θ = 50 °
It has become.
[0032]
FIG. 3 is a block diagram showing a configuration of the reconstruction processing unit 12 shown in FIG.
The reconstruction processing unit 12 includes a reconstruction processing control unit 21, a data memory 22, a back projection block 23, an image memory 24, and an image addition unit 25.
The data memory 22 stores convolution data collected by the data collection unit 11 and subjected to correction processing and convolution processing, as well as division designation data for a target inclined tomographic plane image.
In the first embodiment, since the inclined tomographic plane image is divided into four perpendicular to the Z axis (parallel to the slice plane), the division designation data is the back projection start data for each of the divided areas. The position, backprojection pitch, backprojection start point (addition start point), number of calculations, and the like are stored.
[0033]
The reconstruction processing control unit 21 uses the backprojection block 23 as the control unit body of the reconstruction processing unit 12 to backproject the data stored in the data memory 22 to each voxel. Data obtained by the back projection process is added and stored in the image memory 24 via the image adder 25.
[0034]
The back projection block 23 is independent of the first back projection unit 23-1, the second back projection unit 23-2, the third back projection unit 23-3, and the fourth back projection unit 23-4. And four back projection units capable of parallel processing.
Each of these back projection units 23-1 to 23-4 includes a memory 26, a back projection calculation unit 27, and a data address weight generator 28 as shown in FIG.
[0035]
In the memory 26, the convolution data corresponding to one divided area of the inclined tomographic plane image, the back projection start data position of the divided area, the back projection pitch, the back projection start point (addition start point), the number of calculations (pixels) Number) is transferred and stored.
The backprojection calculation unit 27 uses a data address weight generator 28 based on the backprojection start data position, backprojection pitch, backprojection start point (addition start point), and number of calculations stored in the memory 26. Then, two convolution data are selected from the memory 26 in order, an address is generated, and interpolation is performed by weighting to obtain interpolation data at that address.
[0036]
In the first embodiment, the data interpolation process is performed after the convolution process. However, the interpolation process may be performed on the corrected raw data before the convolution process. .
The backprojection calculation unit 27 multiplies the interpolation data of the convolution data by a predetermined weight and outputs the result to the reconstruction processing control unit 21.
[0037]
In the first embodiment having such a configuration, the reconstruction processing control unit 21 directly reconstructs the inclined tomographic plane image based on the flow of the inclined image reconstruction processing shown in FIG.
[0038]
First, as the processing of step 1 (ST1), an inclined tomographic plane image set from an image obtained by scanogram imaging performed as preliminary imaging is divided into a plurality according to the Z coordinate. For example, as shown in FIG. 6, it is divided into four divided areas Z1, Z2, Z3, and Z4. As for the division method, the division may be performed based on data input by manual operation, or the slice thickness may be set in advance and the slice thickness may be divided in order. is there.
[0039]
Next, as the processing of step 2 (ST2), helical interpolation such as 360 ° interpolation is performed for each divided area on the detection data collected by the data collection unit 11 and subjected to various corrections. Then, data for one rotation (360 °) in the Z coordinate of the divided area is created.
As the process of step 3 (ST3), the data for one rotation created for each divided region in the process of step 2 is subjected to a convolution process.
[0040]
When the creation of the data for one rotation of each divided region is completed, each divided region is assigned to each back projection unit 23-1 to 23-4 as processing in step 4 (ST4), and for each divided region in the processing in step 3. Back-projecting data (back-projection start data position A, back-projection pitch Δ, back-projection start point B, number of calculations (number of pitches), etc.) for each divided area together with the generated convolution processed data It transfers to each memory 26 of the projection parts 23-1 to 23-4.
[0041]
As the processing of step 5 (ST5), as will be described later, direct backprojection processing for each divided region is performed in parallel by the backprojection units 23-1 to 23-4.
When the back projection processing by the back projection units 23-1 to 23-4 is completed, the processing of Step 6 (ST6) is obtained from the back projection units 23-1 to 23-4 as shown in FIG. The image data of each divided area is added in the image memory 24 to complete the data of the inclined tomographic plane image, and this inclined image reconstruction process is completed.
[0042]
FIG. 8 is a diagram showing the flow of back projection processing performed in each of the back projection units 23-1 to 23-4 in the processing of Step 5 described above. Here, the backprojection processing performed by the first backprojection unit 23-1 will be described, and the backprojection processing performed by each of the other backprojection units 23-2 to 23-4 is the same. Is omitted.
First, as the process of step 7 (ST7), the view number (for example, 1000) of the one-dimensional X-ray detector 5 is set in the view number counter formed in the memory 26, and the reconstruction is performed as the process of step 8 (ST8). Based on the data for back projection of the allocated divided area transferred by the processing control unit 21 and stored in the memory 26, the total number of columns in the allocated divided area is set in the column number counter formed in the memory 26. set.
[0043]
Next, in step 9 (ST9), based on the data set in the column number counter and the data stored in the memory 26, the divided area assigned to the pixel number counter formed in the memory 26 is displayed. Set the number of pixels (number of calculations) in the corresponding column.
As the processing of step 10 (ST10), based on the back projection start data position and back projection pitch stored in the memory 26, as shown in FIG. 9 (for easier understanding, in FIG. The back projection processing in one view by the second back projection unit for the divided region is shown, where A is the back projection start data position, B is the back projection start point, Δ is the back projection pitch), and the pixel to be back projected is helical. Two pieces of data subjected to interpolation and convolution processing are selected, weighted interpolation is performed on the two pieces of data by the weight generated by the data address weight generator 28, and this weight is processed as the processing of step 11 (ST11). Multiply the interpolated data by the backprojection weight.
In this back projection process, the interpolation data string from A is once back projected onto the centering line, and the centering process is performed to back project from the centering line to the pixel line from B.
[0044]
In step 12 (ST12), the data multiplied by the backprojection weight is added to the position (address) corresponding to the corresponding pixel in the memory 26.
In step 13 (ST13), it is determined whether or not the count value of the pixel number counter is zero. If it is determined that the count value of the pixel number counter is not 0, a subtraction process of −1 is performed on the count value of the pixel number counter, the back projection process is shifted to the next pixel, and the above-described step 10 is performed again. Return to the process.
[0045]
If it is determined that the count value of the pixel number counter is 0, the back projection process to the one column of the divided region in one view is completed, so that the column number counter is processed as a process of step 15 (ST15). It is determined whether or not the count value is zero. If it is determined that the count value of the column number counter is not 0, a subtraction process of −1 is performed on the count value of the column number counter, the back projection process is shifted to the next column, and the above-described step 9 again. Return to the process.
[0046]
If it is determined that the count value of the column number counter is 0, the backprojection process to the divided area in one view is completed. Therefore, as the process of step 16 (ST16), the count value of the view number counter is It is determined whether or not it is zero. Here, when it is determined that the count value of the view number counter is not 0, the back projection process is shifted to the next view, and the process returns to the above-described step 8 again.
Also, if it is determined that the count value of the view number counter is 0, the backprojection processing to the divided areas in all the views has been completed, and thus the backprojection processing of each backprojection unit ends. Yes.
In each view, for example, as shown in FIG. 10, the backprojection areas required in the respective divided areas are different. FIG. 10 (a) shows the required backprojection area in each view in the first divided area Z1, and FIG. 10 (b) shows the required backprojection area in each view in the second divided area Z2. Show.
[0047]
As described above, according to the first embodiment, the inclined tomographic plane image is collected from the one-dimensional X-ray detector 5 without the voxel data being reconstructed after the voxel data is generated. Can be reconstructed directly from the data.
In this direct reconstruction, the number of back projection operations is
(Total number of pixels in image) x (Number of views)
This is the same as the number of back projection operations in one image of a normal axial section without inclination (slice plane perpendicular to the Z axis). Since this back projection operation accounts for a large proportion (90% in some cases) of the calculation time of image reconstruction processing after data collection, the processing time should be shortened as reconstruction processing of inclined tomographic plane images. Can do.
[0048]
In the first embodiment, a description is given of a case where the number of backprojection units that can be backprojected in parallel with the number of divisions of the inclined tomographic plane image is the same, and parallel backprojection processing is performed for each divided region. However, the present invention is not limited to this. For example, the number of back projection units may be the same as the number of views, and back projection processing may be performed in parallel for each view.
In the first embodiment, the single slice CT apparatus has been described. However, the present invention is not limited to this, and the scan is not a helical scan of a multi-slice CT apparatus using a two-dimensional X-ray detector. The method can also be applied.
[0049]
For example, when a static scan is performed with a 4-row multi-slice CT apparatus, the divided regions of the inclined tomographic plane image shown in FIG. 6 are represented by Z1 = 1st row, Z2 = 2nd row, Z3 = 3rd row, Z4 = If it is formed so as to correspond to the 4th column and “read data of corresponding column” is used instead of each helical interpolation (ST2) in the inclined plane image reconstruction process (see FIG. 5), it is necessary to perform helical interpolation. The sloped tomographic plane image can be directly reconstructed easily.
[0050]
A second embodiment of the present invention will be described with reference to FIGS. In the second embodiment, only the contents of the inclined image reconstruction process in the first embodiment are different, and the basic hardware configuration is the same, so the same members are the same. Reference numerals are assigned and explanations thereof are omitted.
In the second embodiment, the reconstruction processing control unit 21 directly reconstructs the inclined tomographic plane image based on the flow of the inclined image reconstruction processing shown in FIG.
[0051]
First, as the processing of step 21 (ST21), an inclined tomographic plane image set from an image obtained by scanogram imaging performed as preliminary imaging is obtained by converting the Z coordinate perpendicular to the Z coordinate, that is, parallel to the detection slice plane. Disassemble in units.
Next, as a process of step 22 (ST22), a convolution process is performed on the detection data collected by the data collection unit 11 and subjected to various corrections, and as a process of step 23 (ST23), for each Z coordinate. Then, helical interpolation such as 360 ° interpolation is performed to create data for one rotation at each Z coordinate. In this helical interpolation, it is not necessary to perform the data for all the channels obtained from the one-dimensional X-ray detector 5, and if the helical interpolation is performed for the channels in the minimum necessary range, the time required for the interpolation process is reduced. It can be shortened.
[0052]
When the data for one rotation at each Z coordinate has been created, the number of views (for example, 1000) of the one-dimensional X-ray detector 5 is set in the view number counter formed in the data memory 22 as processing in step 24 (ST24). Then, as the processing of step 25 (ST25), the number of columns to be back-projected derived from the Z coordinate range of the inclined tomographic plane image decomposed by the processing of step 21 is set in the column number counter formed in the data memory 22.
[0053]
Next, as the processing of step 26 (ST26), each of the four columns along with the convolution data of the corresponding Z coordinate subjected to helical interpolation corresponding to four sets (one rotation each) corresponding to the next four columns of pixel data. Data for backprojection to pixel data (backprojection start data position A, backprojection pitch Δ, backprojection start point B, number of calculations (number of pitches), etc.) is respectively sent to backprojection units 23-1 to 23-4. Are transferred to the memory 26.
[0054]
As the processing of step 27 (ST27), as will be described later, the backprojection units 23-1 to 23-4 perform direct backprojection processing one column at a time in parallel.
When the backprojection processing by each backprojection unit 23-1 to 23-4 is completed, one column stored in the memory 26 of each backprojection unit 23-1 to 23-4 is processed as step 28 (ST28). Are added to the corresponding column positions in the image memory 24 via the image adder 25.
[0055]
Next, in step 29 (ST29), it is determined whether or not the count value of the column number counter is zero. If it is determined that the count value of the column number counter is not 0, as a process of step 30 (ST30), a subtraction process of -4 is performed on the count value of the column number counter, and the back projection process is performed as the following 4 The process is shifted to the column, and the process returns to step 26 described above.
[0056]
If it is determined that the count value of the column number counter is 0, the back-projection process to the inclined tomographic plane image (pixel) in one view has been completed. Therefore, as the process in step 31 (ST31), the number of views It is determined whether or not the count value of the counter is zero. If it is determined that the count value of the view number counter is not 0, a subtraction process of −1 is performed on the count value of the view number counter, the back projection process is shifted to the next view, and the above-described step 25 is performed again. Return to the process.
If it is determined that the count value of the view number counter is 0, the back-projection process to the inclined tomographic plane image (pixel) in all views is completed, so that this inclined image reconstruction process is ended. It has become.
[0057]
FIG. 12 is a diagram showing the flow of back projection processing performed in each of the back projection units 23-1 to 23-4 in the processing of step 27 described above. Here, the backprojection processing performed by the first backprojection unit 23-1 will be described, and the backprojection processing performed by each of the other backprojection units 23-2 to 23-4 is the same. Is omitted.
First, as the processing of step 33 (ST33), the number of pixels (the number of calculations) in the column corresponding to the pixel number counter formed in the memory 26 is calculated based on the count value of the column number counter and the data stored in the memory 26. set.
[0058]
Next, as the processing of step 34 (ST34), two data subjected to convolution processing and helical interpolation are performed on the pixel to be backprojected based on the backprojection start data position and backprojection pitch stored in the memory 26. Is selected, and weighted interpolation is performed on the two data with the weight generated by the data address weight generator 28. As a process of step 35 (ST35), the weight of back projection is applied to the weighted interpolated data. Multiply
[0059]
As a process of step 36 (ST36), the data multiplied by the back projection weight is added to the position (address) corresponding to the corresponding pixel in the memory 26.
Next, in step 37 (ST37), it is determined whether or not the count value of the pixel number counter is zero. Here, if it is determined that the count value of the pixel number counter is not 0, as a process of step 38 (ST38), a subtraction process of -1 is performed on the count value of the pixel number counter, and the back projection process is performed on the next pixel. The process returns to the above-described step 34 again.
[0060]
If it is determined that the count value of the pixel number counter is 0, the backprojection processing to one column (same Z coordinate) of the inclined tomographic plane image in one view has been completed. The back projection process is terminated. When the back projection process of each back projection unit is completed, the process returns to the above-described tilt image reconstruction process again.
[0061]
As described above, according to the second embodiment, the inclined tomographic plane image can be directly reconstructed with the accuracy of the Z coordinate, and the number of back projection operations in one image of the axial section with almost no inclination is obtained. Therefore, as in the first embodiment, the processing time can be shortened as the reconstruction processing of the inclined tomographic plane image.
In the second embodiment, the single slice CT apparatus has been described. However, the present invention is not limited to this, and can be applied to a multi-slice CT apparatus using a two-dimensional X-ray detector. It is.
[0062]
When applied to this multi-slice CT apparatus, it is also possible to use a filter interpolation method for obtaining data at a predetermined Z coordinate from data having the same phase angle as helical interpolation.
[0063]
Furthermore, in the second embodiment, by performing helical scan interpolation after convolution processing, the number of times of convolution processing is reduced as compared to the case where convolution processing is performed after helical scan interpolation. Thus, the time required for the image reconstruction process can be reduced.
[0064]
In other words, when convolution processing is performed after interpolation of helical scan, data for n slices is created along the Z coordinate by interpolation of helical scan from X-ray detection data for 2 rotations (2 slices). However, these n slices of data are each subjected to convolution processing. However, when interpolation of the helical scan is performed after the convolution processing of the second embodiment, two rotations (two slices) are necessary. X-ray detection data is subjected to convolution processing, and data for n slices is generated along the Z coordinate by interpolation of helical scan from the data for two slices subjected to the convolution processing.
[0065]
Therefore, as the convolution processing, in the second embodiment, two slices are sufficient, but in the method of performing the convolution processing after performing the helical scan interpolation, n slices are necessary.
[0066]
In addition, as a method of performing the convolution processing after performing the helical scan interpolation, a large amount of interpolation data of a necessary phase is created in advance in the Z coordinate direction with a fine slice pitch by the helical scan interpolation, and the corresponding slice It is also possible to select position interpolation data, perform convolution processing on the selected interpolation data, and then backproject to the corresponding pixels.
In this case, in order to create a large amount of interpolation data, helical scan weight data is stored in a table or the like, and when performing helical scan interpolation, the interpolation weight is determined by the slice of the X-ray detection data each time. It is not calculated from the relationship between the position and the desired slice position, but is used by calling from a table or the like.
Further, when this second embodiment is applied to a multi-slice static scan, filter interpolation can be used, and a higher quality image can be obtained by using this filter interpolation.
[0067]
A third embodiment of the present invention will be described with reference to FIGS. Note that the third embodiment is similar to the second embodiment described above, except that the contents of the inclined image reconstruction process in the first embodiment described above are different, and the basic hardware configuration is the same. Since the configuration is the same, the same members are denoted by the same reference numerals and the description thereof is omitted.
In the third embodiment, the reconstruction processing control unit 21 directly reconstructs the inclined tomographic plane image based on the flow of the inclined image reconstruction processing shown in FIG.
[0068]
First, as the processing of step 41 (ST41), an inclined tomographic plane image set from an image obtained by scanogram imaging performed as preliminary imaging is converted to a Z coordinate perpendicular to the Z axis, that is, parallel to the detection slice plane. Disassemble in units.
Next, as a process of step 42 (ST42), a convolution process is performed on the detection data collected by the data collection unit 11 and subjected to various corrections.
[0069]
Next, as the processing of step 43 (ST43), the number of views of the one-dimensional X-ray detector 5 (for example, 1000) is set in the view number counter formed in the data memory 22, and as the processing of step 44 (ST44), The number of columns to be back-projected derived from the Z-coordinate range of the tilted tomographic plane image resolved in step 41 is set in the column number counter formed in the data memory 22.
[0070]
When the setting to each of these counters is completed, as the processing of step 45 (ST45), in order to project back to the pixel data of each 4 columns together with the convolution data for 2 rotations corresponding to the next 4 columns of pixel data. Data (back projection start data position, data rotation position, back projection pitch (channel pitch), data rotation pitch (movement amount of one rotation in the Z coordinate direction), back projection start point, number of calculations (number of pitches), etc.) It transfers to each memory 26 of each back projection part 23-1 to 23-4.
[0071]
Next, as the processing of step 46 (ST46), as will be described later, the backprojection units 23-1 to 23-4 each perform direct backprojection processing in parallel for each column.
When the back projection processing by each back projection unit 23-1 to 23-4 is completed, one column stored in the respective memory 26 of each back projection unit 23-1 to 23-4 is processed as step 47 (ST47). Are added to the corresponding column positions in the image memory 24 via the image adder 25.
[0072]
Next, in step 48 (ST48), it is determined whether or not the count value of the column number counter is zero. If it is determined that the count value of the column number counter is not 0, a subtraction process of −4 is performed on the count value of the column number counter as the process of step 49 (ST49), and the back projection process is performed as the following 4 The process is shifted to the column, and the process returns to step 45 described above.
[0073]
If it is determined that the count value of the column number counter is 0, the back-projection processing to the inclined tomographic plane image (pixel) in one view is completed, so the number of views is determined as the processing in step 50 (ST50). It is determined whether or not the count value of the counter is zero. If it is determined that the count value of the view number counter is not 0, as a process of step 51 (ST51), a subtraction process of -1 is performed on the count value of the view number counter, and the back projection process is performed for the next view. The process returns to the above-described step 44 again.
If it is determined that the count value of the view number counter is 0, the back-projection process to the inclined tomographic plane image (pixel) in all views is completed, so that this inclined image reconstruction process is ended. It has become.
[0074]
FIG. 14 is a diagram showing a flow of back projection processing of each projection unit performed in each of the back projection units 23-1 to 23-4 in the processing of Step 46 described above. Here, the backprojection processing performed by the first backprojection unit 23-1 will be described, and the backprojection processing performed by each of the other backprojection units 23-2 to 23-4 is the same. Is omitted.
First, as the processing of step 52 (ST52), the number of pixels (number of calculation) of the column corresponding to the pixel number counter formed in the memory 26 is calculated based on the count value of the column number counter and the data stored in the memory 26. set.
[0075]
Next, as a process of step 53 (ST53), based on the backprojection start data position Dhj, data rotation position Drj, backprojection pitch Δh, and data rotation pitch Δr stored in the memory 26, the pixel to be backprojected is processed. 4 convolution data (the 2nd data D (m, n), D (m, n + 1) of the nth channel and the n + 1th channel of the corresponding view of the mth rotation, the corresponding view of the m + 1th rotation 2 (D (m + 1, n), D (m + 1, n + 1)) of the n-th channel and the n + 1-th channel are selected, and the data address weight generator 28 is selected for these four data. Weighted interpolation is performed with the weight Wr for helical interpolation (interpolation in the Z-axis direction) and the weight Wh for interpolation in the channel direction.
[0076]
Equation (1) shown below is an equation for calculating weight Wr for helical interpolation, Equation (2) is an equation for calculating Nrk in Equation (1), and Equation (3) is a calculation of weight Wh for correction in the channel direction. The expression (4) is an example of a formula for calculating Nhk in the expression (3).
[0077]
[Expression 1]

[0078]
Next, as the processing in step 54 (ST54), the weighted interpolated data is multiplied by the back projection weight Wb.
[0079]
Here, as the processing of step 55 (ST55), the data multiplied by the back projection weight is added to the position (address) corresponding to the corresponding pixel in the memory 26.
Next, in step 56 (ST56), it is determined whether or not the count value of the pixel number counter is zero. If it is determined that the count value of the pixel number counter is not 0, as a process of step 57 (ST57), a subtraction process of -1 is performed on the count value of the pixel number counter, and the back projection process is performed on the next pixel. The process returns to the above-described step 53 again.
[0080]
If it is determined that the count value of the pixel number counter is 0, the backprojection processing to one column (same Z coordinate) of the inclined tomographic plane image in one view has been completed. The back projection process is terminated.
When the back projection process of each back projection unit is completed, the process returns to the above-described tilt image reconstruction process again.
[0081]
As described above, according to the third embodiment, the same effects as those of the second embodiment described above can be obtained.
In the processing of step 53 and processing of step 54 described above, as shown in FIG. 15, weighted interpolations C1 and C2 of Wh in the channel direction for two data having different phases in the same rotation are performed by parallel processing. If the weighted interpolation C12 of Wr (corresponding to helical scan) in the Z-axis direction is connected to the data obtained by the parallel processing C1 and C2 by pipeline processing, the processing speed can be further improved. it can.
Note that this third embodiment can also be applied to a multi-slice CT apparatus.
[0082]
A fourth embodiment of the present invention will be described with reference to FIGS. Note that, in the fourth embodiment as well, the basic hardware configuration is the same as the previous embodiments, except that the contents of the inclined image reconstruction process in the first embodiment are different. Since they are the same, the same members are denoted by the same reference numerals and description thereof is omitted. In the fourth embodiment, the reconstruction processing control unit 21 directly reconstructs the inclined tomographic plane image based on the flow of the inclined image reconstruction processing shown in FIG.
[0083]
First, as the processing of step 61 (ST61), an inclined tomographic plane image set from an image obtained by scanogram imaging performed as preliminary imaging is decomposed in units of Z coordinates.
Next, as a process of step 62 (ST62), a convolution process is performed on the detection data collected by the data collection unit 11 and subjected to various corrections.
[0084]
Next, as a process of step 63 (ST63), a back projection process with a helical interpolation weight Wr described later is performed, and when the back projection process with this Wr is finished, a process of step 64 (ST64) is performed as (1-Wr When the back projection process according to (1-Wr) is completed, the inclined image reconstruction process is terminated.
[0085]
FIG. 17 is a diagram showing the flow of the back projection process by Wr in the process of Step 62 described above and the back projection process by (1 -Wr) of the process in Step 63 described above.
The backprojection process using (1-Wr) only replaces Wr during backprojection process using Wr with (1-Wr).
First, as the processing of step 65 (ST65), the number of views (for example, 1000) of the one-dimensional X-ray detector 5 is set in the view number counter formed in the data memory 22, and as the processing of step 66 (ST66), The number of columns to be back-projected derived from the Z coordinate range of the tilted tomographic plane image decomposed in step 61 of the tilt image reconstruction process is set in the column number counter formed in the data memory 22.
[0086]
When the setting to each of these counters is completed, the process of step 67 (ST67) is to project back to the pixel data of each 4 columns together with the convolution data for 2 rotations corresponding to the next 4 columns of pixel data. Data (back projection start data position, data rotation position, back projection pitch (channel pitch), data rotation pitch (movement amount of one rotation in the Z coordinate direction), back projection start point, number of calculations (number of pitches, etc.)) It transfers to each memory 26 of each back projection part 23-1 to 23-4.
[0087]
Next, as the processing of step 68 (ST68), back projection processing by Wr (or 1-Wr) of each back projection unit described later is performed, and when the back projection processing by Wr (or 1-Wr) is finished, step 69 (ST69), one column of image data stored in the respective memory 26 of each of the back projection units 23-1 to 23-4 is added to the corresponding column position of the image memory via the image addition unit 25. To do.
[0088]
Next, in step 70 (ST70), it is determined whether or not the count value of the column number counter is zero. If it is determined that the count value of the column number counter is not 0, a subtraction process of −4 is performed on the count value of the column number counter as a process of step 71 (ST71), and the back projection process is performed on the next four columns. The process returns to step 67 described above.
[0089]
If it is determined that the count value of the column number counter is 0, the back-projection process to the inclined tomographic plane image (pixel) in one view has been completed. Therefore, as the process in step 72 (ST72), the number of views It is determined whether or not the count value of the counter is zero. If it is determined that the count value of the view number counter is not 0, as a process of step 73 (ST73), a subtraction process of -1 is performed on the count value of the view number counter, and the back projection process is performed on the next view. Then, the process returns to the above-described step 66 again.
[0090]
If it is determined that the count value of the view number counter is 0, the backprojection processing to the inclined tomographic plane image (pixel) in all the views is completed. Therefore, backprojection by this Wr (or 1-Wr) is performed. The processing is finished. When the back projection processing by Wr (or 1-Wr) is completed, the process returns to the above-described tilt image reconstruction processing again.
[0091]
FIG. 18 is a diagram showing the flow of backprojection processing by Wr (or 1-Wr) of each projection unit performed in each of the backprojection units 23-1 to 23-4 in the processing of Step 68 described above. Here, the backprojection processing performed by the first backprojection unit 23-1 will be described, and the backprojection processing performed by each of the other backprojection units 23-2 to 23-4 is the same. Is omitted.
First, as the processing of step 74 (ST74), the number of pixels in the column corresponding to the pixel number counter formed in the memory 26 is calculated based on the count value of the column number counter and the data stored in the memory 26. set.
[0092]
Next, as a process of step 75 (ST75), two convolutions are performed on the backprojected pixel based on the backprojection start data channel position Dhj, backprojection pitch Δ, and data rotation pitch stored in the memory 26. Data (channel directions generated by the data address weight generator 28 for the data Dk and D (k + 1) of the k-th channel and the k + 1-th channel of the corresponding view at the predetermined number of rotations) In step 76 (ST76), a predetermined backprojection weight Wb and helical interpolation weight Wr (or 1-Wr) are applied to the interpolated data as processing in step 76 (ST76). Multiply products.
In addition, the following formula (6) is a formula for calculating the channel-direction interpolation weight Wh, formula (7) is a formula for calculating Nhk in formula (6), and formula (8) is a formula for calculating the backprojection data Dp. It is an example.
[0093]
[Expression 2]

[0094]
Next, as the processing of step 77 (ST77), the data obtained by multiplying the back projection weight by the helical interpolation weight is added to the position (address) corresponding to the corresponding pixel in the memory 26.
[0095]
Next, in step 78 (ST78), it is determined whether or not the count value of the pixel number counter is zero. If it is determined that the count value of the pixel number counter is not 0, as a process of step 79 (ST79), a subtraction process of -1 is performed on the count value of the pixel number counter, and the back projection process is performed on the next pixel. And the process returns to the process of step 75 described above.
[0096]
If it is determined that the count value of the pixel number counter is 0, the backprojection processing to one column (same Z coordinate) of the inclined tomographic plane image in one view has been completed. The back-projection process using Wr (or 1-Wr) is terminated. When the back projection processing by Wr (or 1-Wr) of each back projection unit is completed, the processing returns to the above-described back projection processing by Wr (or 1-Wr).
[0097]
Therefore, as in the flow of the gradient image reconstruction process shown in FIG. 16, the back projection process by Wr is followed by the back projection process by (1-Wr), thereby achieving helical interpolation. .
As described above, according to the fourth embodiment, the same effects as those of the second embodiment can be obtained.
[0098]
In the fourth embodiment, the back projection process is performed twice to reconstruct the inclined tomographic plane image. However, the present invention is not limited to this, and FIG. As shown, the inclined tomographic plane image may be reconstructed by performing back projection processing n times with helical interpolation weights Wri (i = 1,..., N, Wr1 + Wr2 +... + Wrn = 1). The inclined image reconstruction process shown in FIG. 19 is basically the same as the contents in the inclined image reconstruction process shown in FIG. 16 (see FIGS. 17 and 18), and only the handling of the weight Wr of the helical interpolation is performed. Since they are different, detailed description thereof is omitted here.
[0099]
The back projection process is performed twice (three or more times) by the weight Wr (Wri) of the helical interpolation, but the back projection processes by the weight Wr of the helical interpolation are processed in parallel. If the images are added after the parallel processing, the calculation time can be shortened.
[0100]
By the way, the weight Wr of helical interpolation is set to 1 only in the corresponding area for each area obtained by dividing the inclined tomographic plane image into a plurality according to the Z coordinate as shown in a in FIG. 20, and to 0 in the other areas. If each is weighted, each region is reconstructed separately by reconstructing the inclined tomographic plane image, so that the first embodiment described above can be realized.
[0101]
However, for each area obtained by dividing the inclined tomographic plane image into a plurality of parts according to the Z coordinate as shown in FIG. 20b, the corresponding area is attenuated by 1 and the adjacent areas are attenuated by a linear function. If weighting is performed as described above, in the first embodiment, a step (jagged edge) between regions in one inclined tomographic plane image is improved, and a single inclined tomographic plane image is obtained.
[0102]
Further, as shown in FIG. 21, the helical interpolation weight a in a predetermined view (n) is equal to the helical interpolation weight b in the same phase view (n−N) before one rotation (one rotation is N views). In the portion that does not overlap with the range, the weight is set to 1 so as to attenuate linearly from 1 to 0 in the overlapping portion, and as shown in FIG. 22, it is adjacent to the weight a of a predetermined view (n). Since the weight b in the view (n + 1) is simply a translation in the inclined tomographic plane image, the reconstruction result of the predetermined view (n) can be diverted in the image reconstruction in the adjacent view (n + 1). . Note that the positional relationship in a helical scan of a predetermined view (n), a view (n−N) in phase before one rotation, and an adjacent view (n + 1) is as shown in FIG.
[0103]
Further, since the weight Wr for helical interpolation is a numerical value in a simple format, it may be stored in a table or the like and called at the time of calculation. Further, Wr may be changed in units of pixels to be reconfigured or may be changed in units of pixel columns.
Furthermore, in the fourth embodiment, the helical scan of the single slice CT apparatus has been described. However, the present invention is not limited to this and can be applied to the static scan of the multi-slice CT apparatus. .
[0104]
A fifth embodiment of the present invention will be described with reference to FIGS. As shown in FIG. 24, the fifth embodiment is characterized in that a plurality of adjacent inclined tomographic plane images are reconstructed. Since there is data that can be used redundantly between adjacent inclined tomographic plane images, it is possible to efficiently reconstruct the image. The fifth embodiment is similar to the previous embodiments in that the basic hardware configuration is the same as in the first embodiment except that the contents of the inclined image reconstruction process in the first embodiment are different. Since they are the same, the same members are denoted by the same reference numerals and description thereof is omitted.
[0105]
First, when the pitches of adjacent inclined tomographic plane images are appropriately adjusted, as shown in FIGS. 25 (a) to (d), the inclined tomographic planes are respectively obtained when the adjacent inclined tomographic plane images are divided perpendicularly to the Z axis. All images can be split in the same way.
A first method for reconstructing a plurality of adjacent inclined tomographic plane images is divided as shown in FIGS. 25 (a) to (d), and each is divided as in the first embodiment described above. The back projection process is performed by limiting the inclined tomographic plane image for each divided region. At this time, information such as the back projection start data position, back projection pitch, back projection start point (addition start point), number of calculations, etc. used for back projection for each divided area is common, and calculation processing for handling these information is performed. It can be shortened.
[0106]
In the second method, the first backprojection unit 23-1 to the fourth backprojection unit 23-4 do not perform backprojection processing by accepting one inclined tomographic plane image each of which is in charge. As shown in FIG. 26, each of the back projection units 23-1 to 23-4 takes a part of the same slice position divided perpendicularly to the Z coordinate of the plurality of inclined tomographic plane images.
By doing so, each back projection unit 23-1 to 23-4 does not use the interpolation data redundantly, and one back projection unit uses the interpolation data at one slice position, so that a plurality of inclined slices are used. All portions where the interpolation data existing over the plane image is backprojected are consistently backprojected.
[0107]
The data of each divided region back-projected by each back projection unit 23-1 to 23-4 using dedicated interpolation data is added and synthesized on the image memory 24, and each becomes one inclined tomographic plane image. . In FIG. 26, 11, 21, 31,..., 53, 62, 71 of the inclined tomographic plane image, for example, 11 is the data of the divided region 1 obtained by back projecting the data (interpolation data) 1. Show.
[0108]
Therefore, according to the second method, it is not necessary to create interpolation data for the number of divided regions for each inclined tomographic plane image, and each interpolation data from the data memory 22 corresponds to the corresponding back projection unit 23-1. Since it is only necessary to transfer to one of the memories ˜23-4, it is possible to efficiently reconstruct the image.
As described above, according to the fifth embodiment, it is possible to achieve an efficient image reconstruction capable of reducing the processing time when a plurality of adjacent inclined tomographic plane images are reconstructed.
[0109]
A sixth embodiment of the present invention will be described with reference to FIGS. In the sixth embodiment, a case where static scanning is performed in multi-slice CT will be described. Note that the sixth embodiment also has the same basic hardware configuration as that of the previous embodiments, except that the contents of the tilt image reconstruction process in the first embodiment are different. Since they are the same, the same members are denoted by the same reference numerals and description thereof is omitted. In the sixth embodiment, an example using 4-row multi-slice CT will be described.
[0110]
In the sixth embodiment, the reconstruction processing control unit 21 directly reconstructs an inclined tomographic plane image based on the flow of the inclined image reconstruction processing shown in FIG.
First, as a process of step 81 (ST81), an image obtained by scanogram imaging performed as preliminary imaging is displayed on the display device 14, and an inclined tomographic plane image is set based on a setting operation input. For example, as shown in FIGS. 28 to 30, the set inclined tomographic plane image is divided in accordance with the beam thickness of four rows of X-ray beams perpendicular to the Z coordinate.
[0111]
That is, as shown in FIG. 28, four rows of X-ray beams are radiated to the FOV (imaging region), and the range (each thickness t) of the X-ray beam occupying the central axis of the FOV is shown in FIG. As shown to a), it becomes the position (each thickness t) of each slice perpendicular | vertical to the Z axis | shaft (coinciding with the central axis of FOV) to divide | segment.
As shown in FIG. 29 (a), when an inclined tomographic plane image is set across four slices, it is divided into four regions as shown in FIG. 29 (b). Further, as shown in FIG. 30 (a), when an inclined tomographic plane image is set across three slices, it is divided into three regions as shown in FIG. 30 (b).
[0112]
When the division of the tilted tomographic plane image is completed in this way, as the processing of step 82 (ST82), the gantry is rotated and X-rays are irradiated to collect data for one rotation. Various corrections such as X-ray intensity correction and detector sensitivity correction are performed on the collected data.
When this data collection is completed, the count value 1 is set in the view number counter as the process of step 83 (ST83), and it is determined whether or not the first reconstruction is performed as the process of step 84 (ST84).
[0113]
Here, if it is determined that the reconstruction is the first time, the data in the four columns of the view corresponding to the numerical value set in the view number counter (the data obtained in the processing in step 82) is processed in step 85 (ST85). Are convolved with the reconstruction function to calculate four columns of convolution data.
Next, as the processing of step 86 (ST86), each convolution data is assigned and transferred to each backprojection unit 23-1 to 23-4, and each backprojection unit 23-1 to 23-4 transfers to each divided region. A limited backprojection process is performed, and the image of the divided area obtained by the backprojection process in each of the backprojection units 23-1 to 23-4 is added on the image memory 24 as the process of step 87 (ST87). One inclined tomographic plane image is created and displayed on the display device 14.
[0114]
Next, in step 88 (ST88), it is determined whether or not the count value of the view number counter is equal to the view number N for one rotation.
Here, if it is determined that the count value of the view number counter is not equal to the view number N for one rotation, that is, less than N, as a process of step 89 (ST89), +1 is added to the count value of the view number counter and updated. When the process is completed and the +1 addition update process is completed, the process returns to the above-described step 86 again.
[0115]
Further, when it is determined that the count value of the view number counter is equal to the view number N for one rotation, the display device 14 has an accurate inclined tomographic plane image reconstructed with a necessary number of views (for example, 180 ° may be sufficient). Is displayed. Here, as the processing of step 90 (ST90), it is determined whether or not to update the inclined tomographic plane image.
Here, when it is determined that the inclined tomographic plane image is to be updated, the processing returns to the above-described step 82 again. As a result, in the process of step 84 described above, it is determined that the reconstruction is not the first time but the second time or more.
[0116]
If it is determined that the reconstruction is not the first reconstruction but the second or more reconstruction in the process of step 84 described above, the previous data in the same phase (same view) / detector array is processed as the process of step 91 (ST91). The difference data is obtained by calculating the difference between the data obtained by the collection (corrected data) and the data obtained by the current data collection (corrected data).
[0117]
Next, as the processing of step 92 (ST92), the obtained difference data is convolved with the reconstruction function to obtain difference convolution data. When the difference convolution data is obtained, the process proceeds to the process of step 86 described above, and the difference convolution data is back-projected for each divided region, added in the image memory 24, and the previous inclined fault by the difference convolution data. The surface image is updated (only the changed part of the image is changed).
[0118]
If it is determined in step 90 that the inclined tomographic plane image is not updated, it is determined in step 93 (ST93) whether or not the inclined tomographic plane image is to be changed.
Here, when it is determined that the inclined tomographic plane image is to be changed, the processing returns to the above-described step 81 again.
If it is determined that the inclined tomographic plane image is not changed, the inclined image reconstruction process is terminated.
[0119]
As described above, according to the sixth embodiment, the same effect as that of the first embodiment described above can be obtained in the multi-slice CT static scan.
In the sixth embodiment, an adjacent inclined tomographic plane image may be reconstructed using the technique of the fifth embodiment described above, and the fourth embodiment described above. The inclined tomographic plane image may be reconstructed using a technique such as a form.
[0120]
Further, in order to obtain an image having a large effective slice thickness, for example, as shown in FIG. 31, when 10 rows of two-dimensional X-ray detectors are used, the data of each detector row is added in advance (bundling processing). Just keep it.
That is, at the slice position corresponding to the first detector row, the data of the segment at the same position of the second detector row is added to the doubled data of each segment of the first detector row. In the slice position corresponding to the second detector row, the data of the segments of the first detector row, the second detector row, and the third detector row are added respectively. And at the slice position corresponding to the third detector row, the sum of the data of the segments of the second detector row, the third detector row, and the fourth detector row is used. In the slice position corresponding to the tenth detector row, the data of each segment of the ninth detector row is doubled by the data of the segment in the same position of the tenth detector row. Use what you added .
It should be noted that for the area without data outside the X-ray cone beam, the adjacent data may be approximated by back projection, or may be displayed in black without back projection at all.
[0121]
A seventh embodiment of the present invention will be described with reference to FIGS. 32, 33, 43 and 44. FIG.
In the seventh embodiment, a two-dimensional X-ray detector in which X-ray detection elements are arranged in a planar (two-dimensional) / matrix shape so as to face a conical X-ray (cone beam) is described above. An arbitrary inclined tomographic plane image in the cone beam CT arranged instead of the one-dimensional X-ray detector 5 in the embodiment is reconstructed.
[0122]
The reconstruction method of the two-dimensional image and the three-dimensional image in the cone beam CT is described in detail in Japanese Patent Application No. 8-1015. In Japanese Patent Application No. 8-1015, as shown in FIG. 43, a voxel 203 is set in an imaging region FOV between the X-ray source 201 and the two-dimensional X-ray detector 202, as shown in FIG. Then, after setting the centering surface 204 in which centering rows parallel to the voxel row (for example, 203-1) of the voxel 203 are arranged, the data obtained from the two-dimensional X-ray detector 202 is subjected to interpolation processing or the like, A back projection is performed once on the centering surface 204, and back projection from the centering surface 204 to the voxel 203 is performed to reconstruct a two-dimensional image or a three-dimensional image. With such a reconstruction method, the reconstruction processing time is significantly increased. Shortening can be realized.
[0123]
Further, an arbitrary inclined tomographic plane image can be directly obtained from the X-ray data by reconstructing using the projection method in which the back projection is once performed on the centering plane of Japanese Patent Application No. 8-1015 and then back projected on each voxel row. It is possible to reconfigure.
That is, in the backprojection from the centering plane to the voxels, if only the voxels corresponding to the inclined tomographic plane image are limited in each voxel row, it can be directly reconstructed as in the second embodiment described above. In addition, when the voxel is limited, the centering surface can also be limited. Even in the back projection from the two-dimensional X-ray detector to the centering surface, the back projection process and the result can be re-established as long as the centering surface is limited. The time required for the configuration process can be further shortened.
[0124]
Further, in the cone beam irradiation helical scan, as shown in FIG. 32, when the data obtained from the bifocal beam is back-projected and reconstructed as a tilted tomographic plane image, the tilt image reconstruction shown in FIG. 33 is performed. As in the flow of the configuration process, the inclined tomographic plane image may be divided into regions corresponding to each beam, and reconstructed by two back projection processes in which the regions are limited for each divided region.
[0125]
That is, first, as the processing of step 101 (ST101), 0 ° is set to the phase area θ storing the phase (view), and as the processing of step 102 (ST102), the inclined tomographic plane image is set by the phase area θ. The vertical division is performed on the Z axis corresponding to each focal position in the phase.
[0126]
Next, in step 103 (ST103), an initial value is set in the rotation number counter N, and in step 104 (ST104), the Nth rotation division corresponding to the numerical value set in the rotation number counter N is set. The region-limited cone beam backprojection is performed on the region using the data collected at the phase θ corresponding to the phase set in the phase area θ, and an image reconstructed by backprojection is displayed on the display device 14. When this region-limited cone beam backprojection is completed, it is determined in step 105 (ST105) whether or not the region-limited cone beam backprojection has been completed for all the focal points with respect to the phase θ. That is, it is determined whether or not the cone-beam backprojection on the entire inclined tomographic plane image is completed for the phase θ.
[0127]
If it is determined that the cone-beam backprojection limited to the region is not completed for all the focal points with respect to the phase θ, the process of step 106 (ST106) adds +1 to the count value of the rotation speed counter N. Then, the process returns to the above-described step 104 again.
[0128]
Further, if it is determined that the region-limited cone beam backprojection has been completed for all the focal points with respect to the phase θ, a cone beam backprojection of a necessary phase (360 ° or 180 °) is performed as the processing of step 107 (ST107). It is determined whether or not everything has been completed.
Here, if it is determined that the cone-phase backprojection of the necessary phase has not been completed yet, as a process of step 108 (ST108), an addition update process of + Δθ is performed on the set value of the phase area θ, and the above-described process is again performed The process returns to step 102.
When it is determined that all the necessary cone-beam backprojections have been completed, the tilt image reconstruction process is terminated.
[0129]
As described above, according to the seventh embodiment, the two-dimensional X-ray detector is not reconstructed after creating the voxel data even in the cone beam CT after the voxel data is generated. Can be reconstructed directly from data collected from.
In addition, although the description has been given of the case where the data for one rotation (360 °) is generated by performing the helical interpolation by the 360 ° interpolation method, the present invention is not limited to this. Various interpolation methods such as a filter interpolation method using a beam, an adjacent interpolation method, and an adjacent interpolation method may be used.
[0130]
Furthermore, although the reconstruction method based on the data for one rotation (360 ° + α) is used, the present invention is not limited to this, and the reconstruction method based on the data for half rotation (half, 180 ° + α) is used. You can do it.
Also, the X-ray detector may be 4-row sampling by an equal 8-row detector, a linear multi-slice detector or a flat detector, or data collection by other shapes of X-ray detectors.
In addition, this invention is not limited to embodiment mentioned above, It can deform | transform in the range which does not deviate from the summary of this invention.
[0131]
【The invention's effect】
As described above in detail, according to the present invention, since the inclined tomographic plane image desired by the doctor is divided into a plurality of regions and the reconstruction process is performed, the time required for the reconstruction process of the inclined tomographic plane image is shortened. be able to.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of an X-ray CT apparatus using an image reconstruction method according to a first embodiment of the present invention.
FIG. 2 is a diagram showing a geometry of the X-ray CT apparatus according to the embodiment.
FIG. 3 is a block diagram showing a configuration of a reconstruction processing unit of the X-ray CT apparatus according to the embodiment.
FIG. 4 is a block diagram showing the configuration of a back projection unit of the reconstruction processing unit of the X-ray CT apparatus according to the embodiment.
FIG. 5 is a view showing a flow of a tilt image reconstruction process performed by a reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 6 is a view for explaining division of an inclined tomographic plane image in an inclined image reconstruction process performed by a reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 7 is a diagram for explaining addition of image data of each divided region on the image memory in the tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 8 is a diagram showing the flow of back projection processing in the tilt image reconstruction processing performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 9 is a diagram for explaining a back projection start data position, a back projection start point, and a back projection pitch in a tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 10 is a diagram showing a back projection area required in the tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 11 is a diagram showing a flow of an inclined image reconstruction process performed in a reconstruction processing unit of an X-ray CT apparatus using the image reconstruction method according to the second embodiment of the present invention.
FIG. 12 is a diagram showing the flow of back projection processing of each back projection unit in the tilt image reconstruction processing performed in the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 13 is a diagram showing a flow of an inclined image reconstruction process performed in a reconstruction processing unit of an X-ray CT apparatus using an image reconstruction method according to a third embodiment of the present invention.
FIG. 14 is a diagram showing the flow of back projection processing of each back projection unit in the tilt image reconstruction processing performed in the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 15 is a schematic diagram for explaining weighted interpolation parallel processing and pipeline processing in the gradient image reconstruction processing performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 16 is a diagram showing a flow of an inclined image reconstruction process performed in a reconstruction processing unit of an X-ray CT apparatus using the image reconstruction method according to the fourth embodiment of the present invention.
FIG. 17 is a diagram showing the flow of backprojection processing using Wr (or 1-Wr) in the tilt image reconstruction processing performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 18 shows Wr (or 1-Wr) of each backprojection unit in backprojection processing by Wr (or 1-Wr) in the gradient image reconstruction process performed in the reconstruction processing unit of the X-ray CT apparatus of the embodiment; The figure which shows the flow of the back projection process by).
FIG. 19 is a flowchart showing another example of the tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 20 is a diagram showing an example of weighting for implementing the first embodiment in the tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 21 is a diagram showing weighting for reconstructing a smooth tilted tomographic plane image in the tilted image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 22 is a diagram for explaining diversion of the weighting of FIG. 21 in adjacent views in the tilt image reconstruction processing performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 23 is a diagram showing a positional relationship in a helical scan of a predetermined view, a view of the same phase before one rotation, and an adjacent view of the X-ray CT apparatus of the embodiment;
FIG. 24 shows a plurality of adjacent inclined tomographic images to be reconstructed in a tilted image reconstruction process performed in a reconstruction processing unit of an X-ray CT apparatus using the image reconstruction method of the fifth embodiment of the present invention. The figure which shows a surface image.
FIG. 25 is a view showing a division state of a plurality of adjacent inclined tomographic plane images to be reconstructed in the inclined image reconstruction process of the X-ray CT apparatus of the embodiment.
FIG. 26 is a diagram showing the handling of each back projection unit in the reconstruction of a plurality of adjacent inclined tomographic plane images in the inclined image reconstruction process of the X-ray CT apparatus of the embodiment.
FIG. 27 is a diagram showing a flow of an inclined image reconstruction process performed by a reconstruction processing unit of an X-ray CT apparatus using the image reconstruction method according to the sixth embodiment of the present invention.
FIG. 28 is a diagram showing four rows of cone beams in the X-ray CT apparatus according to the embodiment;
FIG. 29 is a diagram showing a first example of division of four rows of cone beams, inclined tomographic plane images, and inclined tomographic plane images in the X-ray CT apparatus according to the embodiment;
FIG. 30 is a diagram showing a second example of division of four rows of cone beams, an inclined tomographic plane image, and an inclined tomographic plane image in the X-ray CT apparatus according to the embodiment;
FIG. 31 is a diagram for explaining a bundling process when using 10 rows of two-dimensional X-ray detectors in the X-ray CT apparatus according to the embodiment;
FIG. 32 is a diagram showing division of an inclined tomographic plane image with respect to two focal points of a cone beam of an X-ray CT apparatus using the image reconstruction method of the seventh embodiment of the present invention.
FIG. 33 is a diagram showing a flow of a tilt image reconstruction process performed by the reconstruction processing unit of the X-ray CT apparatus according to the embodiment;
FIG. 34 shows a one-dimensional X-ray detector and a two-dimensional X-ray detector of the X-ray CT apparatus.
FIG. 35 is a diagram showing a column direction and a channel direction of the two-dimensional X-ray detector of the X-ray CT apparatus.
FIG. 36 is a diagram for explaining a static scan method and a helical scan method of an X-ray CT apparatus.
FIG. 37 is a diagram showing a rotational phase and a Z-axis position of a view in a helical scan of the X-ray CT apparatus.
FIG. 38 is a diagram showing a slice profile in a helical scan of the X-ray CT apparatus.
FIG. 39 is a diagram showing a cross-sectional image cut out by MPR processing from conventional voxel data.
FIG. 40 is a diagram showing a relationship between pixels of a conventional inclined tomographic plane image and voxel data.
FIG. 41 is a diagram for explaining a jagged state of an image by replacement of voxel data of a conventional inclined tomographic plane image.
FIG. 42 is a diagram showing a designated state of the position of the inclined tomographic plane image in the lumbar scanogram.
FIG. 43 is a view for explaining voxels with respect to a cone beam described in Japanese Patent Application No. 8-1015.
FIG. 44 is a view showing a voxel row and a centering surface described in Japanese Patent Application No. 8-1015.
[Explanation of symbols]
3 ... X-ray source,
5 ... X-ray detector,
11 ... Data collection unit,
12 ... Reconfiguration processing unit,
21 ... Reconfiguration processing control unit,
22: Data memory,
23 ... Backprojection block,
23-1 to 23-4 ... 1st to 4th back projection parts,
24 ... Image memory,
25. Image adding unit,
26 ... Memory,
27: Back projection calculation unit,
28: Data address weight generation unit.

Claims (4)

An X-ray source that generates a fan-shaped X-ray fan beam or a conical X-ray cone beam and an X-ray detector composed of one or more rows of X-ray detection elements continuously rotate around the rotation center axis. In the image reconstruction processing apparatus for collecting helical scan data by moving the top plate on which the subject is placed along the rotation center axis ,
Means for dividing an inclined tomographic plane inclined with respect to the rotation center axis into a plurality of regions along the rotation center axis ;
Means for generating projection data for one rotation for each region from the helical scan data by helical interpolation;
Means for reconstructing a plurality of images related to the plurality of regions by convolution processing and backprojection processing based on the projection data for one rotation ;
Image reconstruction processing apparatus characterized by comprising: means that generates an image corresponding to the inclined tomographic plane from the plurality of images. The image reconstruction processing apparatus according to claim 1, wherein each of the plurality of regions includes a plurality of pixel rows parallel to a slice plane perpendicular to the rotation center axis . When generating projection data for one rotation , after convolution processing is performed on detection data from the X-ray detector, interpolation processing is performed to generate projection data at a corresponding slice position of a pixel group. Item 2. The image reconstruction processing device according to Item 1. When generating projection data for one rotation, the detection data from the X-ray detector is preliminarily interpolated to generate interpolation detection data in the minimum unit of the pitch of the slice position. The image reconstruction processing device according to claim 1, wherein the projection detection data is generated by convolution processing the interpolation detection data.

Images