Computer Generated Holography Using Conjugate Symmetric Extension And Hologram Watermarking

Holography is considered one of the ultimate technologies for three-dimensional display. Compared to conventional methods, computer generated holography (CGH) is advantageous both in generation and reconstruction. CGH has applications in various areas. For example, CGH-based display techniques show a promising future of 3DTV and multimedia presentation. Combination of CGH and digital watermarking provides a novel method of information hiding.As developing effective algorithms is significant to the development of CGH technology, this dissertation first proposes a CGH algorithm based on conjugate symmetric extension and Fourier transform. The complex amplitude of the light wave to be recorded is extended to produce a conjugate symmetric function. The function is then Fourier transformed to generate a real valued distribution containing both amplitude and phase information of the light wave. The obtained real distribution is encoded to give a gray-scale hologram, which can be used to reconstruct the original light wave. Unlike previous algorithms that are based on interference between the object and reference light waves, the proposed method does not need to simulate interference to produce a hologram. The new algorithm is highly efficient since it makes use of the fast Fourier transform. Theoretical derivation is given, and good reconstructions are obtained both digitally and electroholographically. Electroholographic reconstruction is realized by using an electronic addressing reflective LCD spatial light modulator. Experimental results validate the proposed algorithm and show its good performance.In the next part of the dissertation, we study electroholographic reconstruction of color images. CGH algorithms of color images are developed based on the RGB components. Each color component is treated using the method of conjugate symmetric extension and Fourier transform. Digital reconstruction is straightforward, that is, to reconstruct the three color components and combine them to form the color image. In electroholographic reconstruction, however, there is a scaling and alignment problem due to different wavelengths of the three laser sources. Alignment of the reconstructed RGB components is achieved by matching the spatial size and compensating for the spatial offset. Holograms of RGB components are recalculated according to the wavelengths of the light sources, resulting in relevant sizes and offsets. Specifically, spatial spacing of each color component is properly scaled to adjust the spatial frequency of the corresponding hologram so that the reconstructed RGB images can match with each other. Compensation of spatial offset is achieved by controlling reconstruction parameters in the software. A color image is reconstructed using time division switching of reference lights and time division multiplexing of a spatial light modulator.As an application of CGH, we propose a new watermarking technique based on computer generated holography. A hologram of the watermark image is first generated by using the proposed conjugate symmetric extension and Fourier approach. The hologram is embedded into the cover image in the spatial domain. To ensure blind extraction, the spectral components of the embedded hologram and the cover image must be made separable. This is achieved by properly arrange the watermark in an extended mask and perform a slight frequency-domain filtering of the cover image prior to the embedding. The watermarked image has good transparency and robustness. Experiments show that the watermark can survive not only normal image processing such as grayscale transform, smoothing, sharpening, noise contamination, and JPEG compression, but also geometrical attacks including cropping, re-sizing, and rotation. More importantly, the hologram-based watermark is capable of resisting inkjet printing and scanning so that it provides a means of protecting copyright of hardcopy photographs. The proposed watermarking is also useful in detecting tampering as it can reveal details of the geometric transformation of the watermarked image. In addition, we describe an alternative hologram watermarking approach, which uses the quantization index modulation technique to insert a hologram into a cover image in the DCT domain.