A new 2D image compression technique for 3D surface reconstruction

Image compression is one of the important techniques used today for image and video transmission. There are many types of image compression techniques are used these days; one of them is JPEG technique. In this research, we introduce a new idea for applying the JPEG technique with Discrete Wavelet Transform (DWT) for high-resolution images. Our image compression algorithm consists of; firstly, transform an image by single level DWT. Secondly, JPEG algorithm applied on "LL" sub-band this process is called JPEG Transformation. Thirdly, separate the final transformed matrix into DC-Array and AC-Matrix contains DC values and AC coefficients respectively. Finally, the minimize-matrix-size algorithm applied on AC-Matrix followed by arithmetic coding. The novel decompression algorithm used in this research is Parallel Sequential Search Algorithm, which is represented inverse minimize-matrix-size algorithm. The searching algorithm consist of a P pointers, all these pointers are working in parallel to find the original AC-coefficients. Thereafter, combines all decoded DC-values with the decoded ACcoefficients in one matrix followed by apply inverse JPEG transformed and inverse DWT. the technique is tested by compression and reconstruction of 3D surface patches. Additionally, this technique is compared with JPEG and JPEG2000 algorithm by using 2D and 3D RMSE. Keywords—DWT, JPEG, Minimize-Matrix-Size Algorithm, Parallel SS-Algorithm, 3D reconstruction