基於類神經網路模式的電子攝影機色彩恆定及校正

Abstract

在高畫質視訊色彩取像中，我們希望能將電子攝影機所拍攝的影像資訊和 在取像時的原景物有相同的色彩展現。為了達到此一目標，我們將提出與 光源無關且能從彩色影像資訊中重現該物體原有顏色之色彩恆定機制。一 般的電子攝影機使用自動白平衡來補償光源的色偏差，使白色的物體在色 彩重現時仍保持白色。同時儘可能的展現物體的自然色彩。但是自動白平 衡有諸多的條件限制，因此使得色彩重現的品質降低。學者(Worthey)指 出自動白平衡必需要滿足感測器的頻譜響應狹窄並且不相互重疊的條件才 能達到色彩恆定的效果。但這兩個必要條件並無法保証色彩恆定。因為自 動白平衡不能克服感測元件的物理特性不能達到理想頻譜響應中的負響應 部份及光源色相偏紅-綠時所造成的色偏差。為了解決此一問題，我們提 出一個基於非線性類神經網路模式的色彩校正機制可以消除這些非線性的 色彩誤差。最後，結合光源色溫預測及類神經網路模式的色彩校正機制， 我們將可以改進一般傳統電子攝影機的缺點,並可應用於具有色彩恆定功 能的高畫質電子攝影機設計。 In high-fidelity color sensing, one may attempt to use recorded video image data from an electronic camcorder to generate a rendering that appears the same as the original image did at the time of acquisition. To achieve the goal, a color constancy mechanism is proposed to perceive the image color from the strengths of three RGB responses of a color camera or the like, indenpendent of the color of the light illuminating the object. Traditionally, video cameras use the automatic white balance (AWB) to make objects appear as natural as possible under the lighting of the selected scene. It is necessary to compensate for color variations of the lighting so that whites are reproduced as whites. Correct white balance also makes all color correct as well. Nevertheless, there are still several limitations of AWB which would greatly degrade the quality of color reproduction. Worthey showed that the color constancy of AWB is hold when the receptor spectral sensitivities of video cameras should be narrow and do not overlap. But, these two necessary conditions do not imply that the color constancy is achieved because the AWB cannot overcome the physically unrealizable negative lobes involved in the sensitivity curve and it causes the undesired colorimetric shifts when illuminant shifts in the red-green direction. To tackle this difficulty, a nonlinear color correction mechanism based on the feedforward neural network is used to eliminate these nonlinear color error source. The model of a combination of the estimation of the color temperature and the neural-based color correction mechanism can be employed in designing a color constancy electronic video camera which indeed improves the shortcomings of the traditional video camcorders.